Brief Historical Review of Classifications of Asilidae

The eminent dipterist Thomas Becker is quoted in Hermann (1920: 161, see translation below):

Harold Oldroyd, an influential dipterist at the British Museum (Natural History), published the following two comments on classifying Asilidae in 1974:

and

Becker stated that a classification of the genera of Asilidae was needed and that such a project would be worthwhile and rewarding, whereas Oldroyd was pessimistic concerning the results derivable from studying many specimens in detail, comparing their morphology and using character states to propose a hypothesis of relationships. The present study goes against Oldroyd's viewpoint by initiating the study of a representative sample of the 7000 Asilidae species and coding them for 220 morphological characters of the imagines in order to postulate a well-corroborated hypothesis of phylogenetic relationships of higher-level taxa within Asilidae.

The earliest studies dealing with classifications in the form of a dichotomous key were published by Macquart (1838) and Schiner (1862). Hull (1962) summarized the taxonomic work of early authors from the 18th and 19th centuries. Papavero (1973a) provided a detailed summary of the different classifications to 1970 and postulated his own view on the relationships of taxa. Table 1 summarizes classificatory/phylogenetic studies dealing with the entire Asilidae that have been published since 1900. Only a molecular study of 26 Asilidae species (Bybee et al., 2004; fig. 130) can be regarded as a first higher-level phylogenetic analysis of Asilidae involving all previously recognized subfamily taxa (except for Dioctriinae). A few other post-Hennigian publications, published between 1959 and 1996, address the classification of Asilidae as a whole (table 1), but with the exception of Karl (1959; fig. 129) they are not grounded in modern phylogenetic theory. Hull's (1962) and Papavero's (1973a) extensive reviews of the world genera and classification of Asilidae include diagrams of relationships of subordinate taxa. Upon close examination it is apparent that neither of the studies employed a Hennigian character polarization and no sister-group relationships based on apomorphic character states were proposed. Papavero (1973a: 220) cited Darlington (1971) for his general methodology and it is here repeated to illustrate the fact that Papavero did not work in a cladistic framework:

Table 1

Summary of publications addressing the classification or phylogeny of Asilidae based on morphology of imagines except where noted

Darlington (1971: 144) also stated, “In general, I think taxonomists using less arbitrary methods can do as well or better than cladists in fitting their classifications to phylogenies and in interpreting geographical patterns.” In summary, Hull's (1962) and Papavero's (1973a) reviews provided classifications of the genera of Asilidae and identification keys to them, and are thus important contributions to the knowledge of Asilidae, but are not phylogenetic hypotheses of evolutionary relationships of taxa of robber flies. A few phylogenetic hypotheses of relationships of subordinate taxa were published, e.g., Adisoemarto and Wood (1975) on Dioctriini and Echthodopini (Dioctriinae) and Cannings (2002) on Stichopogoninae. Two unpublished Ph.D. dissertations dealt with taxa of Laphriinae, i.e., Andrenosomatini (Fisher, 1986) and the entire Laphriinae (Mazzarolo, 1999).

Taxon Sampling

More than 7000 species have been described in the Asilidae to date, and reflecting this species diversity is of immense importance for any phylogenetic analysis. The present study includes 158 species from 140 genera, representing all 11 previously recognized subfamily taxa with multiple exemplars (table 2), thereby establishing a worldwide sampling (fig. 1). Most of the 42 valid tribal taxa are also represented by multiple exemplars, if applicable. Only the three monogeneric tribal taxa Archilaphriinae Enderlein, 1914 (Dasypogoninae, Archilaphria Enderlein, 1914), Dasylechiini Artigas et al., 1988 (Laphriinae, Dasylechia Williston, 1907), and Neophoneini Artigas et al., 1988 (Laphriinae, Neophoneus Williston, 1889) are not included because too few specimens are available for morphological dissections. To better represent the species diversity additional species are included for every subfamily taxon in which only a few subordinate taxa have been proposed, i.e., Apocleinae, Asilinae, Leptogastrinae, Ommatiinae, Stichopogoninae, and Trigonomiminae. Three species formerly assigned to the Isopogonini Hardy, 1948, which is an unavailable name (Sabrosky, 1999; Dikow, 2004), are included to represent this taxon, which has been applied in previous classifications (e.g., Hardy, 1948; Hull, 1962; Papavero, 1973b). Artigas and Papavero (1997) divided the Asilinae, including the Apocleinae sensu Artigas and Papavero (1988) and Geller-Grimm (2004), into generic groups without formal establishment of family-group names. Seven of these nine generic-group taxa are represented in the present analysis (table 2). The majority of the species are represented by male and female specimens with only three species represented by specimens of the female gender only (table 2). The arrangement of genera in table 2 follows the latest generic classification by Geller-Grimm (2004). An effort is made to represent the included family-group taxa and genera by their respective type genus and type species when material has been available for study. In 11 cases for which the species identification is not certain, i.e., referred to as “sp.,” it has been verified that all specimens belong to a single species and an additional epithet is included in parentheses denoting their collecting locality or unpublished name for future reference. These species are either undescribed, or belong to speciose genera which need to be revised before species identifications can be undertaken.

Table 2

Exemplar species of Asilidae and outgroup taxa in the present analysis Arrangement of genera within subfamily taxa of Asilidae follows Geller-Grimm (2004). Abbreviations: Tg  =  type genus of family-group name; Tsp  =  type species of genus; U  =  E. Fisher unpublished (CSCA); Generic groups sensu Artigas and Papavero (1997): EF  =  Efferia Group, EI  =  Eichoichemus Group, G  =  Glaphyropyga Group, L  =  Lochmorhynchus Group, MA  =  Mallophora Group, MY  =  Myaptex Group, P  =  Proctacanthus Group

Outgroup sampling is also of outstanding importance in a comprehensive phylogenetic analysis. Monophyly of Heterodactyla has been corroborated in a number of analyses (e.g., Yeates, 2002; Yeates and Wiegmann, 2005) and Eremoneura is indisputably a clade (e.g., Grimaldi and Engel, 2005; Yeates and Wiegmann, 2005; Sinclair and Cumming, 2006), so only 16 species representing the Asiloidea have been used as outgroup taxa to polarize the ingroup data (table 2). Neorhynchocephalus volaticus (Williston, 1883) (Nemestrinoidea: Nemestrinidae: Trichopsideinae) is included to represent a taxon outside of Asiloidea and Heterodactyla and is always used as the root of the cladograms presented below. Although the present paper is primarily intended to shed light on the phylogenetic relationships within the Asilidae, it also aims to provide a hypothesis of the sister taxon to Asilidae. Therefore special attention is given to include a number of Apioceridae and Mydidae species, which had historically been discussed as probable sister taxa to Asilidae (see Introduction).

Institutions providing specimens are listed below, together with the abbreviations of the depositories and the people who kindly assisted in lending the specimens:

Phylogenetic Analysis

MacClade (version 4.08, Maddison and Maddison, 2005) is used to enter the data and visualize cladograms found by the parsimony analysis. WinClada (version 0.9.99.88beta, Nixon, 2001) is used to evaluate character optimization/distribution and length of cladograms. The nomenclature for character optimizations applied follows WinClada, i.e., unambiguous, slow ( =  DELTRAN of MacClade), and fast ( =  ACCTRAN of MacClade). The parsimony analysis is conducted in TNT (version 1.0 [July 2004], Goloboff et al., 2001) under UNIX on a 1.33 GHz Apple PowerBook G4 computer. The following provides the commands issued to search for the most parsimonious trees: log ‘file name’; mxram 200; proc ‘file name’; tsave* ‘file name’; hold 150000; mult300; ratchet = iter1500; drift = iter1500; save; nelsen*; save/; log/; (explanation of commands: open log file, set RAM to 200 MB, read matrix, open tree file, hold up to 150000 trees in memory, build 300 random addition sequence Wagner trees and perform TBR branch swapping (RAS+TBR), do 1500 iterations of the parsimony ratchet (Nixon, 1999), do 1500 iterations of tree drifting (Goloboff, 1999), save all trees, do strict consensus, save strict consensus as last tree to file and close tree file, close log file). Bremer support (Bremer, 1988, 1994) and relative Bremer support (Goloboff and Farris, 2001) are calculated in TNT. The Bremer support search is done in an incremental way as described by Giannini and Bertelli (2004) searching for suboptimal trees of one step longer than the most parsimonious tree, then two steps longer, and so on up to eight steps longer, holding an additional 5000 trees for each replicate (commands in script used: macro = ; log ‘file name’; mxram200; proc ‘file name’; tsave* ‘file name’; var 99; set 99 1; loop 1 8 subopt ‘99’; hold+5000; mult300; ratchet = iter1500; drift = iter1500; set 99 ‘99’ + 1; stop save; bsupport = 1; bsupport]; save/; log/; proc/; note: the number 99 is only a name for the variable and the set and does not affect the calculations). This search is started with the command “run ‘file name’” and followed the same exhaustiveness as the searches for most parsimonious trees to sample the tree space as thoroughly as possible for suboptimal topologies. Relative Bremer support takes character conflict into account and provides a proportion as indicated in percent of the number of characters supporting a clade versus the number of characters contradicting the clade (Goloboff and Farris, 2001; Giannini and Bertelli, 2004).

Figure Abbreviations for Morphological Characters

The following abbreviations are found in the figures:

anatg  =  anatergite; anepm  =  anepimeron (mesothoracic); anepst  =  anepisternum (mesothoracic); ant bas  =  anterior basalare; aprn  =  antepronotum; ax cord  =  axillary cord; cerc  =  cercus/i; cerv scl  =  cervical sclerite; cx  =  coxa; d  =  dorsal; d aed shea  =  dorsal aedeagal sheath; ej apod  =  ejaculatory apodeme; epand  =  epandrium; fem  =  femur/femora; goncx  =  gonocoxite; gonst  =  gonostylus/i; hlt  =  halter/es; hypand  =  hypandrium; hyphar  =  hypopharynx; kepm  =  katepimeron (mesothoracic); kepst  =  katepisternum (mesothoracic); ktg  =  katatergite; lat  =  lateral; lat ej proc  =  lateral ejaculatory process; lat pr gonst  =  lateral process/es of gonostyli; lbl  =  labella (labium); lbr  =  labrum; mes  =  mesothoracic; mes pnot  =  mesopostnotum; mes spr  =  mesothoracic spiracle; mes wg  =  mesothoracic wing; met  =  metathoracic; met spr  =  metathoracic spiracle; mr  =  meron; mtanepst  =  metanepisternum; mtepm  =  metepimeron; mtkepst  =  metkatepisternum; mtn  =  metanotum; mx  =  maxilla; plr wg proc  =  pleural wing process; post bas  =  posterior basalare; post ment  =  postmentum (labium); pprn  =  postpronotum; pprn lb  =  postpronotal lobe; pr epand  =  proximal part of epandrium; prepm  =  proepimeron; prepst  =  proepisternum; pre ment  =  prementum (labium); pro  =  prothoracic; sct  =  scutum (mesonotum); sctl  =  scutellum (mesothoracic); sur  =  surstylus/i; trn sut  =  transverse suture; troch  =  trochanter; v aed shea  =  ventral aedeagal sheath.

Head

1. Shape of head: (0) wider than high; (1) circular (as wide as high) (length  =  2; CI  =  0.50; RI  =  0.80). The head of Asilidae is usually wider than high in anterior view (state 0; figs. 26–28). In Broticosia paramonovi (Dioctriinae) and Gonioscelis ventralis, Ospriocerus aeacus, Scleropogon subulatus, and Stenopogon species (all Stenopogoninae), the head is circular in anterior view (state 1; fig. 29). Among the outgroup taxa, the head is always wider than high (state 0).

Figure 26–30

Heads in frontal view. 26. Holcocephala calva (Trigonomiminae). 27. Stichopogon trifasciatus (Stichopogoninae). 28. Clephydroneura sp. (Kepong) (Asilinae). 29. Gonioscelis ventralis (Stenopogoninae). 30. Opomydas townsendi (Mydidae: Ectyphinae). Antennae and vestiture omitted. Scale lines  =  1 mm.

2. Compound eye separation: (0) females and males dichoptic; (1) females dichoptic and males holoptic (length  =  1; CI  =  1.00; RI  =  1.00). In Asilidae both genders possess dichoptic compound eyes (state 0; figs. 26–29). Among the outgroup taxa, the compound eyes of male representatives of Bombyliidae, Nemestrinidae, and Scenopinidae are holoptic (state 1).

3. Vertex shape between compound eyes (in dichoptic females): (0) vertex and compound eyes at same level; (1) vertex slightly depressed (less than 60° angle on median margin of compound eyes); (2) vertex sharply depressed (90° angle on median margin of compound eyes) (length  =  26; CI  =  0.07; RI  =  0.60). Hardy (1927a), Hennig (1973), and Oldroyd (1974a) discussed the depressed vertex in Asilidae and Mydidae. Woodley (1989) postulated the depressed vertex in Mydidae as a possible autapomorphy and concluded that it has arisen independently in robber flies and mydas flies. Here, two different states for the depressed vertex are considered. A slightly depressed vertex is found in the majority of Asilidae (state 1; figs. 26–27). A sharply depressed vertex in which the median margin of the compound eyes is at an angle of 90° is found in Asilinae; most Dioctriinae; Laphriinae; Ommatiinae; and some other species (state 2; fig. 28). Among the outgroup taxa, the vertex is at the same level as the compound eye in most species (state 0), but in Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri, a slightly depressed vertex is developed (state 1; fig. 30).

4. Extent of facial swelling: (0) not developed; (1) lower facial margin slightly protruding; (2) extending over lower facial half; (3) extending over lower three quarters; (4) extending over entire face (length  =  60; CI  =  0.06; RI  =  0.52). Most species of Asilidae have a well-developed facial swelling of varying extent on which the mystax is situated (states 2, 3, 4). In Afroholopogon peregrinus, Heteropogon manicatus, Holopogon priscus, and Metapogon punctipennis (all Brachyrhopalinae) and Trigonomiminae, with the exception of Damalis annulata and Damalis monochaetes, the face is not developed (state 0). In Leptogastrinae except Acronyches maya; Stichopogoninae except Lasiopogon species; Willistonininae except Acnephalum cylindricum; and a few other robber-fly species, only the lower facial margin is slightly protruding (state 1). Among the outgroup taxa, the face extends over the entire height in Mydidae except Neorhaphiomidas sp. (Mandurah) (state 4), but not developed in the remaining outgroup taxa (state 0).

5. Division of face: (0) entire, no division evident; (1) divided, distinct ventral and dorsal halves developed; (2) divided, prominent ventral part and narrow dorsal part developed (length  =  2; CI  =  1.00; RI  =  1.00). The face generally comprises a single sclerite and no transverse subdivision is evident (state 0; fig. 27). In species of Holcocephala (Trigonomiminae) it is divided into a ventral and a dorsal half of equal extent (state 1; fig. 26). Among the outgroup taxa, the face is usually undivided (state 0), but in Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri, a large ventral part and a narrow dorsal part are separated by a transverse indentation (state 2; fig. 30).

6. Mystax: (0) absent; (1) present (length  =  4; CI  =  0.25; RI  =  0.70). The mystax is considered to be an autapomorphic feature of the Asilidae (Oldroyd, 1974a; Woodley, 1989) composed of as little as two setae just dorsal of the fronto-clypeal suture to many setae covering the entire face (see character 7). The only Asilidae species without any mystacal setae is Pseudorus distendens (Dasypogoninae) (state 0). In Mydidae except Neorhaphiomidas sp. (Mandurah), the face is covered with macrosetae very similar in arrangement and extent to assassin flies and these species are therefore coded as possessing a mystax (state 1). This has previously been observed by Hardy (1927a) and Woodley (1989), although neither author homologized the setae in mydas and robber flies. Woodley (1989) stated that the mystax in robber flies is composed of strong macrosetae in contrast to other Diptera with only weak facial setae, but many robber flies, e.g., species of Leptogastrinae and Trigonomiminae, possess only very weak mystacal setae as well. The facial setae in Mydidae are separated medially by an asetose area, with the exception of Mydas clavatus and Rhaphiomidas maehleri, which is never found among Asilidae species.

7. Extent of mystax: (0) restricted to lower facial margin; (1) extending over lower half; (2) extending over lower three quarters, (3) extending over entire face (length  =  60; CI  =  0.05; RI  =  0.46). The extent of the mystacal setae is often associated with the extent of the facial swelling (see character 4), but in a number of species, e.g., Holopogon priscus (Brachyrhopalinae) and Connomyia varipennis (Stenopogoninae), the mystax covers the entire face while the facial swelling is considerably less developed. In some Brachyrhopalinae; many Dasypogoninae; Dioctriinae; Leptogastrinae; Stichopogoninae except Lasiopogon species; and a few other species the mystax is restricted to the lower facial margin (state 0). All remaining character states are found widely within the major taxa of Asilidae. All outgroup taxa lacking a mystax are coded as inapplicable (-).

8. Shape of mystax: (0) setae loosely arranged; (1) setae tightly packed, forming a distinct shield (length  =  2; CI  =  0.50; RI  =  0.80). The mystax is composed of loosely arranged setae that extend anteriorly over the proboscis (state 0). In Lissoteles aquilonius and Stichopogon species (Stichopogoninae) as well as Trichoura sp. (Tierberg) and Willistonina bilineata (Willistonininae), the median setae are closely arranged, forming a shield covering the proboscis (state 1). All outgroup taxa lacking a mystax are coded as inapplicable (-).

9. Clypeus: (0) inverted U-shaped sclerite; (1) inverted U-shaped sclerite, dorsal half sclerotized to form a plate; (2) sclerites extended medially, nearly forming a single plate (length  =  7; CI  =  0.28; RI  =  0.37). In Asilidae, the clypeus appears to be composed of a single plate with a median longitudinal suture (state 2; fig. 32). Among the outgroup taxa, the same configuration is also found in Apioceridae, Apsilocephalidae, Afroleptomydas sp. (Clanwilliam) (Mydidae), Scenopinidae, and Therevidae (state 2). In Poecilognathus sp. (El Hacha) (Bombyliidae), Mitrodetus dentitarsis, Nemomydas brachyrhynchus, Neorhaphiomidas sp. (Mandurah), and Opomydas townsendi (all Mydidae), and Nemestrinidae, the clypeus consists only of an inverted U-shaped sclerite (state 0). Bombylius major (Bombyliidae), Mydas clavatus, Pseudonomoneura hirta, and Rhaphiomidas maehleri (all Mydidae) possess an inverted U-shaped sclerite with the dorsal half forming a plate (state 1).

Figure 31–32

Mouthparts. 31. Head of Apoclea algira (Asilinae) in lateral view. Labrum, maxillae, and hypopharynx protruding from labium; stipites and maxillary palpus grey. For abbreviations see Materials and Methods. 32. Head of Choerades bella (Laphriinae) in frontal view, clypeal sclerites grey. Vestiture omitted. Scale lines  =  1 mm.

10. Clypeal-facial margin (fronto-clypeal suture): (0) forming a distinct ridge; (1) smoothly rounded, not particularly pronounced (length  =  5; CI  =  0.20; RI  =  0.33). The clypeal-facial margin is usually well developed and pronounced in Asilidae (state 0; fig. 31). A smoothly rounded and not particularly pronounced clypeal-facial margin is found in Rhipidocephala sp. (Harold Johnson) and Trigonomima sp. (anamaliensis) (Trigonomiminae) (state 1). Among the outgroup taxa, state 1 is also found in Apsilocephalidae, Nemestrinidae, Scenopinidae, and Therevidae.

11. Maxillary palpus: (0) one-segmented; (1) two-segmented (length  =  17; CI  =  0.05; RI  =  0.81). The number of maxillary palpomeres has been used extensively in Asilidae classifications in the past, e.g., Karl (1959), Hull (1962), Papavero (1973a), and Oldroyd (1974a). In the identification key to taxa provided by Hull (1962), it is used in the first couplet to divide the Asilidae into two large groups highlighting the diagnostic usefulness of this character in separating species of Asilidae. However, Oldroyd (1974a) pointed out that it can be very difficult to count the palpomeres because genal setae often cover the maxillary palpi. Therefore, the heads are macerated here to be able to count the number of palpomeres correctly. One-segmented palpi are found in Asilinae; Bathypogoninae; Andrenosomatini, Cerotainia albipilosa, Ctenotini, Goneccalypsis argenteoviridis, Loewinella nigripes, Orthogonis scapularis, and Smeryngolaphria numitor (all Laphriinae); Leptogastrinae; Ommatiinae; Townsendia albomacula (Stichopogoninae); Trigonomiminae except Rhipidocephala sp. (Harold Johnson); and a few other species (state 0). Two-segmented palpi are found in Brachyrhopalinae; most Dasypogoninae and Dioctriinae; the remaining Laphriinae; Phellinae; Stenopogoninae, with the exception of Daspletis stenoura and Microstylum sp. (Karkloof); and Tillobromatinae (state 1). Among the outgroup taxa, the one-segmented palpi in Mydidae are interpreted as autapomorphic for this taxon by Woodley (1989) (state 0). Apioceridae, Apsilocephalidae, Scenopinidae, and Therevidae have two palpomeres (state 1) whereas Bombyliidae and Nemestrinidae possess only a single palpomere (state 0).

12. Sensory pit in distal palpomere: (0) absent; (1) present (length  =  15; CI  =  0.05; RI  =  0.81). Sensory openings in palpomeres are often found among “nematoceran” Diptera. Yeates (1994) used this character in the Bombyliidae and some species of Asilidae possess a pit in the distal palpomere as well. It is found in one-segmented as well as two-segmented palpi (see character 11). A sensory pit is present in Brachyrhopalinae, with the exception of Afroholopogon peregrinus and Dasycyrton gibbosus; Dasypogoninae, with the exception of Pegesimallus laticornis, Senobasis corsair, Megapoda labiata, and Pseudorus distendens; Echthodopa pubera (Dioctriinae); Hoplistomerus nobilis, Psilocurus modestus, Protometer sp. (El Tuparro), and Trichardis effrena (all Laphriinae); Leptogastrinae; Stenopogoninae; with the exception of Daspletis stenoura and Microstylum sp. (Karkloof), Stichopogoninae; Tillobromatinae; Trigonomima sp. (anamaliensis) (Trigonomiminae); and Willistonininae (state 1). Among the outgroup taxa, sensory pits are found in Apioceridae, Apsilocephalidae, Bombylius major (Bombyliidae), Opomydas townsendi (Mydidae), Scenopinidae, and Therevidae (state 1). E. Fisher (personal commun.) pointed out in his review of this publication that Laphriinae (except Andrenosomatini and possibly Atomosiini) always possess a sensory pit in contrast to my observations. In this study, the heads have been macerated to be able to examine the palpi in detail; moreover, no SEM micrographs were taken, which would have increased the chances of detecting a pit. I do not believe, although I cannot claim to know the truth, that my possible miscodings of a few Laphriinae species (and possibly other species) would change the cladistic relationships of the studied Asilidae species. I, therefore, suggest that future studies pay particular attention to this character and use SEM micrographs to verify the presence/absence of pits in the maxillary palpi.

13. Distal palpomere: (0) cylindrical; (1) laterally compressed (leaflike) (length  =  2; CI  =  0.50; RI  =  0.83). The distal palpomere is usually cylindrical in Asilidae (state 0; fig. 31). Fisher (1986) employed the laterally compressed distal palpomere in species of Andrenosomatini (Laphriinae) as an autapomorphic feature. State 1 is found in all Andrenosomatini and Apioceridae and all remaining taxa have a simple, cylindrical distal palpomere (state 0).

14. Stipites fusion: (0) fused entirely medially; (1) fused medially, but with V-shaped indentation; (2) divided medially; (3) not sclerotized (length  =  35; CI  =  0.08; RI  =  0.62). The stipites of the maxilla are two separated sclerites bearing the maxillary palpi in Diptera (McAlpine, 1981). In the Asilidae, the stipites are large platelike sclerites at the base of the proboscis (fig. 31) that can be fused medially or divided medially. In most Brachyrhopalinae, in Dasypogoninae, Leptogastrinae, and Trigonomiminae, and in some other species of various taxa the stipites are fused medially and form a single sclerite (state 0; fig. 34). In many Asilinae and Laphriinae; Ommatiinae; Connomyia varipennis, Creolestes nigribarbis, Daspletis stenoura, Microstylum sp. (Karkloof), and Prolepsis tristis (all Stenopogoninae); Tillobromatinae; and a few other species, the two stipites are fused only partly and a V-shaped indentation is developed medially (state 1). In some Asilinae; Bathypogoninae; Dioctriinae; Laphriinae; Gonioscelis ventralis, Ospriocerus aeacus, Scleropogon subulatus, and Stenopogon species (all Stenopogoninae); Stichopogoninae; and other species the stipites are divided medially and a longitudinal fissure is evident that ends at the fusion line with the postgenae (state 2; fig. 33). Among the outgroup taxa, some Mydidae and Therevidae possess divided stipites (state 2) and the stipites in Nemestrinidae are not sclerotized (state 3).

Figure 33–34

Heads in posterior view. 33. Stichopogon trifasciatus (Stichopogoninae). 34. Apoclea algira (Asilinae). Occipital foramina grey; vestiture omitted. Scale lines  =  1 mm.

15. Stipites fusion to postgenae: (0) separated; (1) fused (length  =  3; CI  =  0.33; RI  =  0.86). The medioventral area of the posterior head of Diptera is formed by a sclerotized hypostomal bridge and called the postgenae (McAlpine, 1981). In Asilidae except Holcocephala calva (Trigonomiminae), the stipites are fused to the postgenae and only the lateral apices are free (state 1; figs. 33–34). Among the outgroup taxa, only Apioceridae also possesses fused stipites and postgenae (state 1) and in all other species the stipites are separated from the postgenae (state 0).

16. Medial projection of postgenae: (0) absent; (1) present (length  =  2; CI  =  0.50; RI  =  0.66). In general, the postgenae are flat medially and this state is found in all Asilidae (state 0). Among the outgroup taxa, Apioceridae and Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri (Mydidae) possess a median projection on the postgenae (state 1).

17. Shape of proboscis: (0) straight; (1) angled slightly and evenly ventrally; (2) angled sharply ventrally and posteriorly; (3) distal tip angled slightly and evenly dorsally (length  =  5; CI  =  0.60; RI  =  0.33). Most Asilidae species possess a more or less straight proboscis (state 0; fig. 31), but in a few species, the proboscis is angled dorsally or ventrally. An evenly ventrally angled proboscis is found in Archilestris magnificus and Megapoda labiata (Dasypogoninae) (state 1) and the distal tip is evenly angled dorsally in Dasyllis funebris, Pilica formidolosa, and Proagonistes praeceps (Laphriinae) (state 3). Ancylorhynchus fulvicollis (Stenopogoninae) is unique in that the proboscis is sharply angled ventrally and slightly posteriorly (state 2). Oldroyd (1974b) discussed the function and morphology of the proboscis in Ancylorhynchus in detail. Among the outgroup taxa, all species possess a straight proboscis (state 0).

18. Cross section of prementum proximally: (0) circular; (1) dorsoventrally flattened; (2) laterally compressed; (3) square; (4) not sclerotized (length  =  14; CI  =  0.28; RI  =  0.78). The sclerotized labium in Asilidae forms a ventral sheath surrounding the maxillae and hypopharynx (fig. 31). In contrast to Crampton (1942: fig. 2F) and Wood (1981: 550), I consider the proximal part of the labium to be the postmentum and not the prementum (see fig. 31). In most Asilidae the prementum is more or less circular in diameter (state 0). A dorsoventrally flattened prementum is found in Lycomya germainii (Asilinae) and Andrenosomatini (Laphriinae) (state 1). A laterally compressed prementum is most pronounced in all Laphriini (Laphriinae; fig. 32) in which the proboscis appears knifelike, but the prementum is also laterally compressed although to a lesser degree in Ceraturgus fasciatus, Chrysopogon pilosifacies, and Codula limbipennis (all Brachyrhopalinae); Dasypogoninae, with the exception of Dasypogon diadema, Lestomyia fraudiger, Saropogon species, and Threnia carbonaria; Daspletis stenoura and Microstylum sp. (Karkloof) (Stenopogoninae); Hypenetes stigmatias (Tillobromatinae); and Trigonomiminae, with the exception of Rhipidocephala sp. (Harold Johnson) and Trigonomima sp. (anamaliensis) (state 2). A square cross section is found in Dioctria species and Nannodioctria sp. (formicaphaga) (Dioctriinae) (state 3). Among the outgroup taxa, an unsclerotized prementum without a distinct cross section is found in most species (state 4), with the exception of Bombyliidae, Rhaphiomidas maehleri (Mydidae), and Nemestrinidae, which have a circular prementum (state 0).

19. Ventral development of postmentum: (0) ventral margin entirely smooth; (1) median groove only in distal part; (2) median groove on entire length (length  =  25; CI  =  0.08; RI  =  0.57). The postmentum, the proximal part of the labium (fig. 31), is usually smooth on its ventral surface and appears platelike in Asilidae (state 0). A median groove is developed in a number of species and the extent of this groove is variable. A short median groove restricted to the distal postmentum is present in Blepharotes splendidissimus (Asilinae); Bathypogoninae; Ceraturgus fasciatus, Nannocyrtopogon lestomyiformis, Nicocles politus, Metapogon punctipennis, and Rhabdogaster pedion (all Brachyrhopalinae); Blepharepium cajennensis (Dasypogoninae); Dioctriinae except Broticosia paramonovi; Andrenosoma cyrtoxys, Dasyllis funebris, Hoplistomerus nobilis, Perasis transvaalensis, Proagonistes praeceps, and Smeryngolaphria numitor (all Laphriinae); Phellinae; Stenopogoninae, with the exception of Stenopogonini; Stichopogon punctum (Stichopogoninae); and Willistonininae (state 1). A long median groove extending over the entire length of the postmentum is found in Austrosaropogon nigrinus, Cophura brevicornis, Dasycyrton gibbosus, and Holopogon priscus (all Brachyrhopalinae) and most Dasypogoninae (state 2). Among the outgroup taxa, state 2 is also found in Therevidae. All remaining outgroup taxa possess a smooth ventral margin (state 0).

20. Ventral margin of postmentum: (0) platelike, straight; (1) platelike, arched (length  =  1; CI  =  1.00; RI  =  1.00). The ventral margin of the postmentum is usually straight in Asilidae (state 0; fig. 36) and only Dioctria species and Nannodioctria sp. (formicaphaga) (Dioctriinae) possess an arched postmentum (state 1; fig. 35). Among the outgroup taxa, the postmentum is always platelike and straight (state 0).

Figure 35–38

Labia and maxillary palpus and lacinia in lateral view. 35. Dioctria atricapillus (Dioctriinae) labium. 36. Choerades bella (Laphriinae) labium. 37. Colepia rufiventris (Asilinae) lacinia. 38. Diogmites grossus (Dasypogoninae) lacinia. Vestiture omitted. Scale lines  =  1 mm.

21. Post- and prementum fusion: (0) not fused; (1) entirely fused (length  =  4; CI  =  0.25; RI  =  0.81). The post- and prementum are generally separated by a membranous area and appear not entirely fused in most Asilidae (state 0; fig. 35). In Megapoda labiata, Pseudorus distendens, and Senobasis corsair (all Dasypogoninae), and Andrenosomatini, Ctenotini, Laphriini, Hoplistomerus nobilis, and Perasis transvaalensis (all Laphriinae), the post- and prementum are entirely fused (state 1; fig. 36). Among the outgroup taxa, post- and prementum are never fused (state 0).

22. Dorsal flange medially on prementum: (0) absent; (1) present (length  =  22; CI  =  0.04; RI  =  0.50). The two median, dorsal margins of the prementum can be elevated and form a flange that can be seen easily in lateral view. Most species of Asilidae lack such a median flange, i.e., Asilinae except Proctacanthus philadelphicus; Laphriinae; Leptogastrinae; Ommatiinae; Stichopogoninae except Lasiopogon aldrichii; Trigonomiminae except Damalis sp. (Palatupana); and some other species (state 0). Many Brachyrhopalinae; Dasypogoninae; Stenopogoninae; Willistonininae; and some other species possess a dorsal flange (state 1). Among the outgroup taxa, a dorsal flange is never developed (state 0).

23. Labella: (0) well developed and separated from prementum; (1) reduced, fused to prementum only ventrally; (2) reduced, fused to prementum entirely (length  =  13; CI  =  0.15; RI  =  0.84). In Diptera the labella are generally enlarged lobes at the tip of the proboscis that function to take up liquids, yeasts, bacteria, spores, pollen, or other fine particles for nutrition. Hennig (1973) and Woodley (1989) recognized the reduced labella and the fusion to the prementum as an autapomorphy of Asilidae related to the predatory habit of the imagines. Upon close examination, it becomes apparent that the labella are not always entirely fused to the prementum, which has been highlighted for Dioctriinae species by Adisoemarto and Wood (1975), who used the presence/absence of a “subapical notch” dividing the labella dorsally from the prementum as a character. In Asilinae; Bathypogoninae; most Brachyrhopalinae; Ommatiinae; Phellinae; Stichopogoninae; and Willistonininae; and most Stenopogoninae, the labella are only partly fused to the prementum (state 1; fig. 31). In Chrysopogon pilosifacies and Codula limbipennis (Brachyrhopalinae); Dasypogoninae, with the exception of Dasypogon diadema, Lestomyia fraudiger, Saropogon luteus, and Thereutria amaraca; Laphriinae, with the exception of Laphystia species, Laxenecera albicincta, Nusa infumata, Orthogonis scapularis, Protometer sp. (El Tuparro), Psilocurus modestus, Smeryngolaphria numitor, Trichardis effrena, and Zabrops tagax; Leptogastrinae except Acronyches maya; Daspletis stenoura and Microstylum sp. (Karkloof) (Stenopogoninae); Hypenetes stigmatias and Lycostommyia albifacies (Tillobromatinae); and Trigonomiminae, the labella are entirely fused to the prementum (state 2; figs. 35–36). Among the outgroup taxa, the labella and prementum are always separated (state 0).

24. Apex of labella: (0) rounded; (1) pointed; (2) blunt; (3) enlarged, pseudotracheae present (length  =  23; CI  =  0.13; RI  =  0.76). In the Asilidae the sclerotized labella can be of varying shape. In most species the tip is simple and rounded, i.e., in most Asilinae; Bathypogoninae; Leptogastrinae; Ommatiinae; Phellinae; and Willistonininae (state 0; fig. 31). A pointed proboscis is found in most Brachyrhopalinae; Dasypogoninae; Stenopogoninae; and Trigonomiminae (state 1). Within Laphriinae, states (0), (1), and (2) are found: Andrenosomatini, Atomosia puella, and Perasis transvaalensis possess a pointed proboscis (state 1), Ctenotini, Cerotainia albipilosa, Laphriini, Orthogonis scapularis, and Smeryngolaphria numitor possess a blunt tip (state 2; fig. 36; see also McAlpine, 1981; fig. 54), and the remaining taxa possess a rounded tip (state 0). Among the outgroup taxa, broadened labella with pseudotracheae are developed (state 3).

25. Length of labella: (0) small, only forming distal tip of proboscis; (1) large, occupying from one-third to one-half the length of proboscis (length  =  25; CI  =  0.04; RI  =  0.71). The length of the labella varies considerably in Asilidae and two character states are recognized here. Small labella, which only form the distal tip of the proboscis, are found in Asilinae, with the exception of Asilus crabroniformis, Blepharotes splendidissimus, Proctacanthus philadelphicus, and Satanas gigas; Bathypogoninae; many Brachyrhopalinae; Leptogastrinae; Ommatiinae; Stichopogoninae; and Willistonininae, with the exception of Acnephalum cylindricum and Sisyrnodytes sp. (Gamka) (state 0; fig. 31). Large labella are found in Dasypogoninae except Saropogon luteus; Dioctriinae except Broticosia paramonovi; most Laphriinae; Phellinae; Stenopogoninae; Tillobromatinae; and Trigonomiminae (state 1; figs. 35–36). Among the outgroup taxa, both character states are present.

26. Extent of sensory setae on prementum and labella: (0) setae on prementum and labella; (1) setae restricted to labella (length  = 11; CI  =  0.09; RI  =  0.52). The labella and sometimes also the prementum possess sensory setae on the anterior, dorsal, and ventral surfaces and the extent of these setae varies among Asilidae. In general, the setae are restricted to the labella (state 1), but a few species, i.e., Cabasa pulchella, Ceraturgus fasciatus, Cophura brevicornis, Heteropogon manicatus, Holopogon priscus, Nannocyrtopogon lestomyiformis, Nicocles politus, and Rhabdogaster pedion (all Brachyrhopalinae); Archilestris magnificus and Saropogon luteus (Dasypogoninae); Broticosia paramonovi and Myelaphus melas (Dioctriinae); most Laphriinae; and Coleomyia setigera, possess setae on the labella and the prementum (state 0). All species in which the labella and prementum are entirely fused (see character 23) are coded as inapplicable (-) because no decision can be made concerning the border between the structures. Among the outgroup taxa, sensory setae are only developed on the labella (state 1).

27. Lacinia length: (0) shorter than labrum; (1) longer than labrum (length  =  3; CI  =  0.33; RI  =  0.77). I follow Hennig (1973) and McAlpine (1981) in naming the maxillary blade ( =  endite) of the maxilla the lacinia. Martin (1968) stated that the maxillae of Leptogastrinae are only very poorly developed, a finding not supported here. The length of the lacinia does not vary among species of Asilidae, where it is always longer than the labrum (state 1), but in the outgroup taxa. Mydidae, Scenopinidae, and Hemigephyra atra (Therevidae) possess a shortened lacinia that is shorter than the labrum (state 0).

28. Shape of lacinia (lateral view): (0) same height throughout; (1) proximally high, distal tip distinctly narrower: (2) tapered gradually toward tip (length  =  18; CI  =  0.11; RI  =  0.74). Melin (1923) illustrated the laciniae of a number of robber flies and it can be seen that the shape of the laciniae varies greatly within Asilidae species. The laciniae are of equal height throughout and only the very distal tip might be narrowed as in Asilinae, with the exception of Blepharotes splendidissimus and Proctacanthus philadelphicus; Bathypogoninae; Dioctriinae; most Laphriinae; Leptogastrinae; Ommatiinae; most Stenopogoninae; Trigonomiminae; Tillobromatinae except Lycostommyia albifacies; and Willistonininae (state 0; fig. 37). A proximally high, distinctly narrowed lacinia distally with a dorsal notch is found in most Brachyrhopalinae; Dasypogoninae; Andrenosomatini except Dasyllis funebris and Ctenotini (all Laphriinae); Phellinae; and a few Stenopogoninae (state 1; fig. 38). A gradually tapered lacinia is found in Blepharotes splendidissimus (Asilinae); Dasyllis funebris (Laphriinae); and Acnephalum cylindricum (Willistonininae) (state 2). Among the outgroup taxa, state 2 is also found in Apsilocephalidae, Mydidae, Nemestrinidae, and Phycus frommeri (Therevidae) whereas all remaining species possess state 0. It is sometimes difficult to distinguish between states (1) and (2).

29. Tip of lacinia: (0) blunt; (1) pointed (length  =  4; CI  =  0.25; RI  =  0.91). The tip of the lacinia is generally pointed in Asilidae (state 1; figs. 37–38), but in Asilinae, with the exception of Asilus crabroniformis and Colepia rufiventris; Ommatiinae; and Willistonina bilineata (Willistonininae) the tip is blunt (state 0; fig. 31). Among the outgroup taxa, blunt laciniae occur (state 0).

30. Hypopharynx: (0) only weakly sclerotized; (1) heavily sclerotized (length  =  1; CI  =  1.00; RI  =  1.00). The unpaired hypopharynx lies between the labrum and labium as a tonguelike structure. It is generally weakly sclerotized in Diptera, as is here found in all outgroup taxa (state 0), and more heavily sclerotized in blood sucking flies, e.g., Culicidae, Tabanidae, and Pupiparia, and predatory flies. Woodley (1989) recognized the heavily sclerotized, needlelike hypopharynx in all robber flies as an autapomorphy (state 1; figs. 39–41).

Figure 39–43

Hypopharynces in dorsal and lateral view. 39. Colepia rufiventris (Asilinae) dorsal. 40. Acronyches maya (Leptogastrinae) dorsal. 41. Archilestris magnificus (Dasypogoninae) dorsal. 42. Stenopogon rufibarbis (Stenopogoninae) tip in detail (lateral). 43. Acronyches maya tip in detail (lateral). Spicules omitted from 39–41. Scale lines  =  1 mm.

31. Hypopharynx length: (0) short, shorter than labium; (1) long, as long as labium (length  =  2; CI  =  0.50; RI  =  0.92). Yeates (1994) discussed the different lengths of the hypopharynx in several Asiloidea taxa and pointed out that it is relatively longer in Asilidae. Asilidae is coded to possess a long hypopharynx (state 1; fig. 31). Among the outgroup taxa, the hypopharynx is as long as the labium in Bombylius major (Bombyliidae) and Nemestrinidae (state 1) whereas it is shorter than the labium in all remaining taxa (state 0).

32. Shape of sclerotized hypopharynx: (0) parallel-sided throughout, only distal tip suddenly pointed; (1) parallel-sided in proximal half, distal half much narrower; (2) tapering gradually toward distal tip (length  =  15; CI  =  0.13; RI  =  0.76). The shape of the hypopharynx varies within Asilidae and most species have a parallel-sided hypopharynx with a suddenly pointed tip, i.e., Asilinae; Bathypogoninae; Dioctriinae, with the exception of Broticosia paramonovi and Myelaphus melas; Laphriinae; Leptogastrinae; Ommatiinae; most Stenopogoninae; Stichopogoninae; and Trigonomiminae (state 0; figs. 39–40). A proximally parallel-sided and distally gradually tapering hypopharynx is found in most Brachyrhopalinae; Dasypogoninae; Phellinae; Tillobromatinae; and a few other species (state 1; fig. 41). Among the outgroup taxa, most species possess a gradually tapering hypopharynx (state 2), with the exception of Apioceridae and Nemestrinidae, which exhibit state 0.

33. Hypopharynx dorsal protuberances: (0) absent; (1) present, seta-like spicules spaced close together; (2) present, seta-like spicules spaced far apart (length  =  4; CI  =  0.50; RI  =  0.91). Melin (1923) illustrated the strong seta-like spicules on the dorsal margin of the hypopharynx and called it “bristly covering.” These spicules are only present in Asilidae, but two different developments are apparent. In the majority of species, they are spaced close together (state 1; fig. 42), but in Leptogastrinae; Plesiomma sp. (Guanacaste) (Stenopogoninae); and Trichoura sp. (Tierberg) (Willistonininae) these spicules are spaced far apart (state 2; fig. 43). Among the outgroup taxa, spicules are absent (state 0).

34. Length of labrum: (0) long, as long as labium (or nearly so); (1) short, at most half the length of labium (length  =  1; CI  =  1.00; RI  =  1.00). All Asilidae possess a short labrum that is at most half as long as the labium (state 1; fig. 31) whereas all outgroup taxa possess a long labrum (state 0).

35. Dorsoposterior margin of cibarium: (0) simple, no particular ridge connecting cornua; (1) one transverse ridge connecting cornua, cornua originating from ridge; (2) one transverse ridge connecting cornua, cornua not originating from ridge (length  =  9; CI  =  0.22; RI  =  0.89). Yeates (1994) pioneered the study of internal head structures across Asiloidea by illustrating the cibarium and tentorium for many representatives. The cibarium functions as a food pump and is associated with the hypopharynx (Snodgrass, 1944; McAlpine, 1981). Its shape varies within Asilidae and the transverse ridges between the cornua (see character 37) are absent only in Ctenotini (Laphriinae); Cophinopoda chinensis (Ommatiinae); Gonioscelis ventralis (Stenopogoninae); and Willistonina bilineata (Willistonininae) (state 0; fig. 46). A single transverse ridge from which the cornua originate is present in Brachyrhopalinae; Dasypogoninae; Dioctriinae; Laphriinae except Ctenotini; Acronyches maya (Leptogastrinae); Phellinae; Stenopogoninae except Stenopogonini; Stichopogoninae; Tillobromatinae; and Trigonomiminae (state 1; fig. 47). In Asilinae; Bathypogoninae; the remaining Leptogastrinae; Ommatiinae; and Stenopogoninae a single transverse ridge is found, but the cornua do not originate from this ridge (state 2). Among the outgroup taxa, no transverse ridges are developed (state 0).

Figure 44–47

Cibarium. 44–45. Cibarium with attached pharyngeal pump in lateral view (dotted lines indicate unsclerotized cuticle connecting both structures, anterior  =  left). 44. Dioctria hyalipennis (Dioctriinae). 45. Lycomya germainii (Asilinae). 46–47. Cibarium in posterior view. 46. Willistonina bilineata (Willistonininae). 47. Dioctria hyalipennis. Scale lines  =  0.5 mm.

36. Shape of cibarium: (0) trapezoidal; (1) smoothly rounded; (2) dorsally narrow and ventrally broad (length  =  2; CI  =  1.00; RI  =  1.00). Yeates (1994) found that the shape of the cibarium varies within Asiloidea and all Asilidae species he studied possess a trapezoidal cibarium. This is confirmed here and only Asilidae have a trapezoidal cibarium (state 0; figs. 46–47). Among the outgroup taxa, all species possess a smoothly rounded cibarium (state 1), with the exception of Mydas clavatus (Mydidae), which has a dorsally narrow and ventrally broad cibarium (state 2; see figs. in Yeates, 1994).

37. Development of cornua of cibarium: (0) well developed in anteroposterior orientation; (1) well developed in dorsoventral orientation; (2) reduced or absent (length  =  12; CI  =  0.16; RI  =  0.82). Melin (1923) illustrated the cibarium (referred to as pharynx) in lateral view, highlighting the cornua pharyngis, i.e., the dorsolateral attachment of the protractor muscles, which can be oriented differently. Following Owsley (1946), these muscle attachment projections are here simply termed cornua. Martin (1968) stated that all Leptogastrinae possess only reduced cornua, a finding not supported here. Yeates (1994) called the cornua “posterior arms of cibarium” and discussed the different orientation in Bombyliidae. In Asilidae, the orientation varies as well and in Asilinae; Holopogon priscus (Brachyrhopalinae); Loewinella nigripes (Laphriinae); Leptogastrinae, with the exception of Acronyches maya and Beameromyia bifida; Ommatiinae, with the exception of Cophinopoda chinensis and Ommatius tibialis; Stenopogonini except Gonioscelis ventralis (Stenopogoninae); and Stichopogoninae the cornua are oriented in an anterior-posterior fashion (state 0; fig. 45). In the remaining Asilidae and outgroup taxa the cornua are oriented in a dorsoventral fashion (state 1; fig. 44). In Beameromyia bifida (Leptogastrinae); Mydas clavatus (Mydidae), and Nemestrinidae, the cornua are reduced or absent (state 2).

38. Median longitudinal ridge on cibarium: (0) absent; (1) present (length  =  2; CI  =  0.50; RI  =  0.93). Yeates (1994) mentioned that all Asilidae studied possess a median longitudinal ridge on the cibarium. This ridge is found here in all Asilidae and Neorhaphiomidas sp. (Mandurah) (Mydidae) (state 1; figs. 46–47), but absent in all remaining outgroup taxa (state 0).

39. Pharyngeal pump: (0) absent; (1) present (length  =  3; CI  =  0.33; RI  =  0.71). Melin (1923) referred to the pumping structure dorsal to the cibarium as the swollen esophagus, which is later termed pharyngeal pump by Owsley (1946). The anterior part of this pump is situated between or just dorsal of the cornua (Owsley, 1946). It is present in all Asilidae and most outgroup taxa (state 1; figs. 44–45), with the exception of Bombyliidae and some Mydidae species (state 0).

40. Pharyngeal pump development: (0) laterally without sclerotized ring; (1) laterally with sclerotized ring, but medio-anteriorly unsclerotized; (2) with entirely sclerotized ring (length  =  11; CI  =  0.18; RI  =  0.75). The lateral margin of the pharyngeal pump can be unsclerotized or possess a sclerotized ring of varying extent. The lateral margin is unsclerotized in Phellinae; Stenopogoninae, with the exception of Creolestes nigribarbis and Plesiomma sp. (Guanacaste); Tillobromatinae; and Willistonininae except Ablautus coquilletti (state 0). A sclerotized ring that is not entirely closed medio-anteriorly is found in the majority of Asilidae and all outgroup taxa (state 1; fig. 49). An entirely sclerotized ring is present in Cyrtophrys attenuatus, Deromyia fuscipennis, Megapoda labiata, and Pseudorus distendens (all Dasypogoninae); Atomosiini except Smeryngolaphria numitor and Laphriini (Laphriinae); as well as Plesiomma sp. (Guanacaste) (Stenopogoninae) (state 2; fig. 48). All species not possessing a pharyngeal pump (see character 39) are coded as inapplicable (-).

Figure 48–49

Pharyngeal pump in dorsal view. 48. Orthogonis scapularis (Laphriinae). 49. Dioctria hyalipennis (Dioctriinae). Scale lines  =  0.5 mm.

41. Pharyngeal pump shape (dorsal view): (0) circular; (1) triangular (length  =  10; CI  =  0.10; RI  =  0.47). In dorsal view, the pharyngeal pump is generally circular (state 0; figs. 48–49), but in a few species it is more or less of triangular shape, i.e., Dasypogon diadema and Lestomyia fraudiger (Dasypogoninae); Phellinae; Stenopogoninae, with the exception of Ancylorhynchus fulvicollis, Gonioscelis ventralis, Plesiomma sp. (Guanacaste), and Prolepsis tristis; Tillobromatinae except Hypenetes stigmatias; and Sisyrnodytes sp. (Gamka) (Willistonininae) (state 1). Among the outgroup taxa, Apsilocephalidae is the only taxon with a triangular pharyngeal pump (state 1). All species not possessing a pharyngeal pump (see character 39) are coded as inapplicable (-).

42. Dorsal margin of postocciput: (0) simple, no projections; (1) triangular projection present (length  =  3; CI  =  0.33; RI  =  0.93). The dorsal margin of the postocciput is usually not particularly pronounced although in some instances a narrow ridge is developed (state 0; fig. 51). In Dasycyrton gibbosus (Brachyrhopalinae); Myelaphus melas (Dioctriinae); and Laphriinae, a prominent triangular projection is present that is best seen in dorsal view (state 1; fig. 50). Among the outgroup taxa, all species exhibit state 0.

Figure 50–51

Heads in dorsal view. 50. Dasyllis funebris (Laphriinae). 51. Pegesimallus laticornis (Dasypogoninae). Vestiture omitted except for setae on median occipital sclerite. Scale lines  =  1 mm.

43. Shape of occipital foramen: (0) a single foramen; (1) two apparent foramina divided by a bridge (length  =  1; CI  =  1.00; RI  =  1.00). Yeates (1994) highlighted the variation in development of the occipital foramina in Asiloidea and here the majority of species possess a single foramen (state 0; figs. 33–34). Two distinct foramina are present only in Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri (state 1).

44. Anterior tentorial pits: (0) small, slitlike, inconspicuous, ventrally located; (1) well developed, conspicuous, anteroventrally located; (2) well developed, conspicuous, anteromedially located; (3) absent (not observable) (length  =  5; CI  =  0.60; RI  =  0.84). Papavero (1973a) used the development of the anterior tentorial pits (referred to as deep, mediogenal grooves), as a diagnostic character for the Trigonomiminae. In most Asilidae species, the tentorial pits are small, slitlike, and inconspicuously positioned more or less ventrally (state 0; figs. 27–29, 32). In all Trigonomiminae, they are well developed, conspicuous, and positioned anteroventrally on the head (state 1; fig. 26). In Phellus olgae (Phellinae) and Plesiomma sp. (Guanacaste) (Stenopogoninae), the tentorial pits are positioned much more dorsal on the head and this position is coded as a separate character state (state 2). Among the outgroup taxa, state 0 is primarily found, but state 2 is observed in Mydidae (fig. 30), with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri, and in Poecilognathus sp. (El Hacha) (Bombyliidae) the anterior tentorial pits appear to be absent (state 3).

45. Facial setae: (0) only circular setae; (1) circular and dorsoventrally flattened setae laterally in males and females; (2) circular and dorsoventrally flattened mystacal setae in males; (3) circular and dorsoventrally flattened mystacal setae in males and females (length  =  9; CI  =  0.33; RI  =  0.25). In Asilidae, the well-developed facial setae are usually only circular (state 0), but a few species possess setae of different cross section. The facial setae are sometimes differently developed among female and male specimens of a single species. Circular as well as dorsoventrally flattened setae that are positioned laterally to the mystax in males and females are found only in Laphriini except Laphria flava (Laphriinae) (state 1). Circular and dorsoventrally flattened mystacal setae are present in males only of Goneccalypsis argenteoviridis, Orthogonis scapularis, and Perasis transvaalensis (all Laphriinae) (state 2). The same mystacal setae are present in females and males of Smeryngolaphria numitor (Laphriinae); Lissoteles aquilonius (Stichopogoninae); and Ablautus coquilletti and Sisyrnodytes sp. (Gamka) (Willistonininae) (state 3). Pseudorus distendens (Dasypogoninae), which does not possess any facial setae, is coded as inapplicable (-). Among the outgroup taxa, only circular setae are developed (state 0).

46. Size of anterior median ommatidia: (0) all same size; (1) at least some median ommatidia distinctly larger (length  =  26; CI  =  0.03; RI  =  0.64). As early as Walker (1851), dipterists realized that the median ommatidia of the flattened compound eyes of an assassin fly are often enlarged. Although it is sometimes difficult to determine whether the median ommatidia are larger than surrounding lateral ones, I attempted to code this character. The majority of Asilidae species possess larger ommatidia medially (state 1) with notable exceptions of all Bathypogoninae; Phellinae; and Stenopogoninae, in which all ommatidia are of the same size (state 0). Among the outgroup taxa, all ommatidia are of the same size (state 0).

47. Postgenae: (0) posterior margin simple, smoothly rounded; (1) posterior margin with large flangelike projection (length  =  1; CI  =  1.00; RI  =  1.00). The posterior margin of the postgenae is generally smoothly rounded in Asilidae and all outgroup taxa (state 0). In Lamyra gulo and Stiphrolamyra angularis (Laphriinae) it is developed as a flangelike projection. Londt (1988) used this character (referred to as posterior margin of lower occiput) in an identification key to Afrotropical Laphriinae diagnosing the two genera. See character 16. Among the outgroup taxa, the postgenae are smoothly rounded (state 0).

48. Shape of frons (at level of antennal insertion): (0) more or less parallel-sided; (1) markedly and suddenly diverging laterally; (2) markedly approximating medially (length  =  5; CI  =  0.40; RI  =  0.70). The frons is usually parallel-sided as wide at the ventral margin as at the level of antennal insertion in Asilidae (state 0; figs. 26, 28–29). In a few species, i.e., Cerotainia albipilosa (Laphriinae); Stichopogoninae; as well as Trichoura sp. (Tierberg) and Willistonina bilineata (Willistonininae), the frons is markedly and suddenly diverging laterally (state 1; fig. 27). The opposite case is found in Acronyches maya (Leptogastrinae) and Plesiomma sp. (Guanacaste) (Stenopogoninae), in which the compound eyes approximate each other markedly at the level of the antennal insertion (state 2). Among the outgroup taxa, a parallel-sided frons is present (state 0).

49. Macrosetae on lateral margin of frons, dorsal to antennal insertions: (0) absent; (1) present (length  =  30; CI  =  0.03; RI  =  0.40). Asilidae usually possess setae on the lateral margin of the frons dorsal to the antennal insertions. These setae can be particularly well developed and form strong macrosetae. Regular setae are present in Brachyrhopalinae; Leptogastrinae; Ommatiinae; Tillobromatinae; Trigonomiminae; and Willistonininae (state 0) whereas all remaining major taxa have at least some representatives possessing stronger macrosetae (state 1). Among the outgroup taxa, no macrosetae are developed (state 0).

50. Longitudinal ridge medially on frons, dorsal to antennal insertion: (0) absent; (1) present, connecting antennae and ocellar triangle (length  =  2; CI  =  0.50; RI  =  0.50). A longitudinal ridge medially on the frons is developed in Cerotainia albipilosa, Goneccalypsis argenteoviridis, and Loewinella nigripes (all Laphriinae) (state 1) whereas it is absent in all remaining Asilidae (state 0). Among the outgroup taxa, a longitudinal ridge is absent (state 0).

51. Position of antennae on head: (0) in ventral half; (1) in dorsal half (length  =  5; CI  =  0.20; RI  =  0.42). The position of the antennae on the head varies within Asiloidea. In all Asilidae, the antennae are positioned in the dorsal half of the head (state 1; figs. 26–29). Among the outgroup taxa, in Apsilocephalidae, Bombylius major (Bombyliidae), and Mydidae except Neorhaphiomidas sp. (Mandurah) the antennae are positioned in the dorsal half (state 1), but in the remaining taxa, the antennae originate in the ventral half of the head (state 0).

52. Distance between scapices: (0) scapices close together, touching; (1) scapices separated (length  =  2; CI  =  0.50; RI  =  0.83). Yeates and Irwin (1996) established the very closely associated scapices as an autapomorphy of Mydidae. In the present analysis, only Mydidae except Neorhaphiomidas sp. (Mandurah) possess closely associated scapices (state 0; fig. 30) and all remaining taxa possess separated scapices (state 1; figs. 26–29).

53. Elevation of antennal insertion: (0) not elevated; (1) elevated above eye margin in lateral view, with distinct protuberance (length  =  8; CI  =  0.12; RI  =  0.61). The antennal insertion is usually not visible in lateral view on a robber-fly head. A few species possess a protuberance on which the antennae originate and the base of the scape is therefore visible in lateral view, i.e., Austrosaropogon nigrinus and Cabasa pulchella (Brachyrhopalinae); Cyrtophrys attenuatus and Saropogon luteus (Dasypogoninae); Dioctriinae except Broticosia paramonovi; Cerotainia albipilosa and Laphriini (Laphriinae); and Ospriocerus aeacus (Stenopogoninae) (state 1). Among the outgroup taxa, the antennal insertion is not elevated (state 0).

54. Shape of postpedicel: (0) tapering distally; (1) cylindrical, same diameter throughout; (2) medially broadest, long; (3) elongated and distinctly clubbed distally; (4) proximal bulb with elongated distal part; (5) medially broadest, short (length  =  44; CI  =  0.11; RI  =  0.66). In Brachycera the postpedicel is situated distal to the pedicel. In Asilidae and other “orthorrhaphous” taxa, it is “a compound structure formed by sequential distal fusion of the third segment with successive distal segments” (Stuckenberg, 1999: 41) of varying length. All parts distal to the postpedicel are considered here to comprise the stylus (see characters 57–58) (Stuckenberg, 1999). Contrary to the view of Hennig (1972: fig. 29), the postpedicel of Mydidae is considered to entail the cylindrical proximal part and the expanded distal part, which are separated by an unsclerotized part (see character 56). The shape of the postpedicel varies considerably in Asilidae and five character states (fig. 52) are employed here to account for this morphological diversity. The character-state distribution is as follows: Asilinae: state 1, Lycomya germainii; state 2, Threnia carbonaria; state 5, Dasophrys crenulatus, Eichoichemus pyrrhomystax, Lochmorhynchus albicans, Philodicus tenuipes, Pogonioefferia pogonias, Proctacanthus philadelphicus, and Promachus amastrus; state 0, all remaining species; Bathypogoninae, state 2; Brachyrhopalinae: state 0, Dasycyrton gibbosus, Holopogon priscus, and Heteropogon manicatus; state 2, Brachyrhopala ruficornis and Cabasa pulchella; state 1, all remaining species; Dasypogoninae: state 1, Archilestris magnificus, Cyrtophrys attenuatus, and Dasypogon diadema; state 2, all remaining species; Dioctriinae: state 2, Eudioctria albius and Nannodioctria sp. (formicaphaga); state 1, all remaining species; Laphriinae: state 1, Atomosiini, Hoplistomerus nobilis, Laphystia aegyptiaca, and Perasis transvaalensis; state 2, all remaining species; Leptogastrinae: state 1, Acronyches maya; state 0, all remaining species; Ommatiinae: state 0, Michotamia aurata; state 5, all remaining species; Phellinae: state 1, Phellus olgae; state 2, Obelophorus terebratus; Stenopogoninae: state 0, Scleropogon subulatus and Stenopogon species; state 2, Creolestes nigribarbis, Daspletis stenoura, and Microstylum sp. (Karkloof); state 1, all remaining species; Stichopogoninae: state 1, Lasiopogon cinctus and Townsendia albomacula; state 5, Lissoteles aquilonius; state 0, all remaining species; Tillobromatinae: state 1, Lycostommyia albifacies; state 2, all remaining species; Trigonomiminae: state 4, Damalis species; state 1, all remaining species; Willistonininae: state 2, Ablautus coquilletti and Willistonina bilineata; state 0, all remaining species. Outgroup taxa: state 0, Apsilocephalidae, Bombylius major (Bombyliidae), and Scenopinidae; state 1, Poecilognathus sp. (El Hacha) and Therevidae; state 2, Rhaphiomidas maehleri (Mydidae); state 3, Mydidae except other mentioned species; state 5, Apioceridae and Neorhaphiomidas sp. (Mandurah) (Mydidae).

Figure 52

Schematic drawing of shape of postpedicel coded in character 54 with character states indicated. Character state 3, which is only found in Mydidae, not illustrated; vestiture omitted; not to scale.

55. Setae dorsally on postpedicel: (0) absent; (1) present (length  =  32; CI  =  0.03; RI  =  0.63). Many species of Asilidae possess setae on the dorsal surface of the postpedicel (state 1; fig. 53). Setae are absent in Bathypogoninae; Leptogastrinae; Phellinae; Tillobromatinae; and several other species (state 0). Lampria clavipes (Laphriinae) and Ospriocerus aeacus (Stenopogoninae) are coded polymorphic (0/1), as individuals with and without setae are examined. Among the outgroup taxa, setae are present only in Apioceridae, Poecilognathus sp. (El Hacha) (Bombyliidae), as well as Neorhaphiomidas sp. (Mandurah), Opomydas townsendi, and Rhaphiomidas maehleri (all Mydidae) (state 1).

Figure 53–57

SEM micrographs of antennae in lateral view. 53. Dysmachus trigonus (Asilinae) postpedicel. 54. Philonicus albiceps (Asilinae) stylus and enlarged apical “seta-like” sensory element. 55. Dioctria atricapillus (Dioctriinae) postpedicel and stylus. 56. Damalis femoralis Ricardo, 1925 (Trigonomiminae), base of postpedicel with apical “seta-like” sensory element. 57. Hoplistomerus nobilis (Laphriinae) tip of postpedicel with apical “seta-like” sensory element. Scale lines  =  100 µm except for inset scale line  =  10 µm.

56. Sclerotization of postpedicel: (0) entirely sclerotized; (1) cylindrical proximal part and bulbous distal part separated by unsclerotized area (length  =  1; CI  =  1.00; RI  =  1.00). Asiloidea species usually possess an entirely sclerotized postpedicel (state 0), but in Mydidae, with the exception Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri, an unsclerotized median part is developed (state 1, see character 54).

57. Position of apical ‘seta-like’ sensory element: (0) apically on stylus; (1) apically in cavity on stylus; (2) subapically in cavity on postpedicel; (3) subapically and laterally on stylus; (4) apically in cavity on postpedicel; (5) subapically and laterally on postpedicel (length  =  31; CI  =  0.16; RI  =  0.69). Papavero (1973a) discussed the development and position of the stylus in Asilidae in detail. The nomenclature pertaining to these parts is adjusted here to make more meaningful homology statements, however. The stylus in “orthorrhaphous” Brachycera ( =  part of flagellum sensu McAlpine, 1981) was defined by Stuckenberg (1999: 41) as, “the structure consisting of any free segments distal to the pedicel, hypothesized to terminate apically in segment 10.” Hennig (1972) called the terminal element of the stylus that corresponds to the tip of segment 10 Endgriffel, and this element was termed apical “seta-like” sensory element by Dikow and Londt (2000). As the stylus is fused to the postpedicel in stepwise fashion (Stuckenberg, 1999, see characters 54 and 58), only the apical “seta-like” sensory element is differentiable and its position on the stylus or postpedicel varies greatly within Asilidae. In Asilinae; Bathypogoninae; most Brachyrhopalinae; Leptogastrinae; Ommatiinae; most Willistonininae; and several other species the apical “seta-like” element is situated apically on the stylus (state 0; figs. 53–54). In Austrosaropogon nigrinus and Cyrtopogon rattus (Brachyrhopalinae); Dasypogon diadema, Lestomyia fraudiger, Megapoda labiata, Pegesimallus laticornis, and Saropogon species (all Dasypogoninae); Hoplistomerus nobilis, Laphystia species, Nusa infumata, Perasis transvaalensis, Protometer sp. (El Tuparro), Psilocurus modestus, Trichardis effrena, and Zabrops tagax (all Laphriinae); Phellus olgae (Phellinae); Ancylorhynchus fulvicollis, Ospriocerus aeacus, Plesiomma sp. (Guanacaste), and Scylaticus costalis (all Stenopogoninae); Tillobromatinae; Willistonina bilineata (Willistonininae); and Coleomyia setigera the apical “seta-like” element is situated apically on the stylus, but in a distinct apical cavity (state 1; fig. 57). In Chrysopogon pilosifacies and Codula limbipennis (Brachyrhopalinae); Deromyia fuscipennis (Dasypogoninae); and Damalis species (Trigonomiminae), the apical “seta-like” element is situated in a subapical cavity on the postpedicel (state 2, e.g.; fig. 56). In Ceraturgus fasciatus (Brachyrhopalinae); Dioctriinae; Townsendia albomacula (Stichopogoninae); and Rhipidocephala sp. (Harold Johnson) (Trigonomiminae); the apical “seta-like” element is situated subapically laterally on the stylus (state 3; fig. 55). A “seta-like” element is situated in an apical cavity on the postpedicel in Brachyrhopala ruficornis and Cabasa pulchella (Brachyrhopalinae); most Dasypogoninae; Andrenosomatini, Ctenotini, Laphriini, and Laxenecera albicincta (all Laphriinae); Daspletis stenoura and Prolepsis tristis (Stenopogoninae); and Trigonomima sp. (anamaliensis) (Trigonomiminae) (state 4). It is situated subapically and laterally on the postpedicel in Cyrtophrys attenuatus (Dasypogoninae) and Atomosiini (Laphriinae) (state 5). Among the outgroup taxa, Hemigephyra atra (Therevidae) and Scenopinidae exhibit state 1 and Poecilognathus sp. (El Hacha) (Bombyliidae) and Mydidae exhibit state 4 whereas all remaining taxa possess an apical “seta-like” sensory element apically on the stylus (state 0).

58. Stylus: (0) three elements; (1) two elements; (2) one element; (3) reduced, only apical “seta-like” sensory element present (length  =  33; CI  =  0.09; RI  =  0.68). The stylus (see characters 54 and 57) in Asilidae is generally composed of one or two elements (states 1 and 2). In Brachyrhopala ruficornis, Cabasa pulchella, Chrysopogon pilosifacies, and Codula limbipennis (all Brachyrhopalinae); most Dasypogoninae; Andrenosomatini, Atomosiini, Ctenotini, Laphriini, and Laxenecera albicincta (all Laphriinae); Daspletis stenoura and Prolepsis tristis (Stenopogoninae); and Damalis species and Trigonomima sp. (anamaliensis) (Trigonomiminae), the stylus is only represented by the apical “seta-like” sensory element (state 3). Among the outgroup taxa, a three-segmented stylus is only present in Nemestrinidae (state 0) and the remaining states are also represented in the outgroup taxa.

59. Setae on stylus: (0) absent; (1) present on anterior surface only; (2) present on all surfaces (length  =  2; CI  =  1.00; RI  =  1.00). Hardy (1927a), in erecting the tribe Ommatiini, pointed out the setose antennae of species of Ommatius Wiedemann, 1821 and other genera. All Ommatiinae are found here to possess setae on the anterior surface of the stylus (state 1) and Oligopogon sp. (Cape Recife) possesses setae on all surfaces of the stylus (state 2). Among the outgroup taxa, all species lack setae on the stylus (state 0). All species, in which the stylus is reduced to the apical “seta-like” element (see character 58), are coded as inapplicable (-).

60. Ocelli on ocellar triangle: (0) three ocelli on a single ocellar triangle; (1) anterior ocellus separate, situated anteriorly (length  =  1; CI  =  1.00; RI  =  1.00). In Diptera the three ocelli are usually situated on a single ocellar triangle (McAlpine, 1981) and this is found in all Asilidae (state 0; figs. 26–29, 50–51). Among the outgroup taxa, in Apioceridae and Mydidae, the anterior ocelli are separated and are situated more anteriorly (state 1; fig. 30) whereas they are situated on a single ocellar triangle in the remaining species (state 0).

61. Development of posterior ocelli: (0) circular or slightly oval; (1) distinctly elongated (length  =  1; CI  =  1.00; RI  =  1.00). Yeates and Irwin (1996) highlighted the distinct elongation of the posterior ocelli in Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri, and this is confirmed here (state 1; fig. 30). All remaining taxa have three equally large, circular to slightly oval ocelli (state 0; figs. 26–29, 50–51).

62. Ocellar (oc) setae: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  36; CI  =  0.05; RI  =  0.60). Hardy (1930) described the development of ocellar setae in Australian robber flies in detail. Most Asilidae species possess setae or macrosetae on the ocellar triangle (states 1 and 2), but they are found to be absent in Leptogastrinae (state 0). Among the outgroup taxa, there are either no setae (state 0) or only regular setae developed (state 1).

63. Shape of postcranium: (0) flat to convex; (1) concave (length  =  2; CI  =  0.50; RI  =  0.87). Yeates (1994) observed a distinctly concave postcranium in many Bombyliidae. In the present analysis, a concave postcranium is found in Bombyliidae, Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri, and Nemestrinidae (state 1). All Asilidae and remaining outgroup taxa possess a flat to slightly convex postcranium (state 0).

64. Shape of posterior compound eye margin (lateral view): (0) straight or slightly curved anteriorly throughout; (1) sinuate in ventral quarter; (2) distinctly sinuate in ventral half (length  =  7; CI  =  0.28; RI  =  0.85). The majority of Asilidae have a more or less oval-shaped compound eye in lateral view (state 0; fig. 58). In most Asilinae, with the exception of Blepharotes splendidissimus, Clephydroneura species, Heligmonevra laevis, Lycomya germainii, and Satanas gigas; as well as Lestomyia fraudiger (Dasypogoninae); and Willistonininae, with the exception of Acnephalum cylindricum and Sisyrnodytes sp. (Gamka), the ventral quarter is angled anteriorly and the ventral part of the eye is much narrower than the dorsal half (state 1; fig. 59). All Stichopogoninae are outstanding in that the ventral half is much narrower than the dorsal half and the compound eye appears sinuate in lateral view (state 2; fig. 60). Among the outgroup taxa, all species possess state 0.

Figure 58–60

Compound eyes in lateral view. 58. Pegesimallus laticornis (Dasypogoninae). 59. Pogonioefferia pogonias (Asilinae). 60. Stichopogon trifasciatus (Stichopogoninae). Anterior  =  left. Scale lines  =  1 mm.

65. Macrosetae on median occipital sclerite (m ocp scl): (0) absent; (1) two distinct macrosetae; (2) more than two macrosetae (length  =  10; CI  =  0.20; RI  =  0.55). The median occipital sclerite is situated ventral to the vertex and between the compound eyes and often entirely fused to the occiput (McAlpine, 1981). In Dasypogoninae, the sclerite is relatively easy to distinguish and in most instances, more than two macrosetae are situated on it (state 2). In Archilestris magnificus and Pseudorus distendens, no setae are developed (state 0; fig. 50) and in Cyrtophrys attenuatus, Deromyia fuscipennis, Pegesimallus laticornis, and Senobasis corsair, only two macrosetae are present (all Dasypogoninae) (state 1; fig. 51). In Heteropogon manicatus (Brachyrhopalinae); Hyperechia nigripennis (Laphriinae); Daspletis stenoura (Stenopogoninae); Lycostommyia albifacies (Tillobromatinae); and Acnephalum cylindricum and Sisyrnodytes sp. (Gamka) (Willistonininae), more than two macrosetae are developed as well (state 2). Among the outgroup taxa, all species lack any macrosetae (state 0).

66. Postocular (pocl) setae: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  17; CI  =  0.11; RI  =  0.42). Postocular setae, often also labeled occipital setae in robber flies, are present in all Asilidae studied and most species possess macrosetae and regular setae (state 2). Only regular setae are developed in Phellinae; Trigonomiminae, with the exception of Damalis monochaetes; and a few Brachyrhopalinae; Leptogastrinae; Dasypogoninae; and other species (state 1). Among the outgroup taxa, all three character states are observed.

67. Shape of postocular (pocl) setae: (0) straight or only slightly angled anteriorly; (1) sharply angled anteriorly in distal half (length  =  12; CI  =  0.08; RI  =  0.26). The postocular setae are usually straight or only slightly angled anteriorly in Asilidae (state 0). In Dasophrys crenulatus, Neoitamus cyanurus, and Neolophonotus bimaculatus (all Asilinae); Nannocyrtopogon lestomyiformis (Brachyrhopalinae); Dicolonus simplex and Echthodopa pubera (Dioctriinae); Laphria species (Laphriinae); Emphysomera species and Ommatius incurvatus (Ommatiinae); Creolestes nigribarbis, Gonioscelis ventralis, and Stenopogon rufibarbis (all Stenopogoninae); and Lasiopogon species (Stichopogoninae), the setae are long and sharply angled anteriorly in the distal half (state 1). Among the outgroup taxa, all species possess straight postocular setae (state 0).

Thorax

For reference; figure 61 illustrates the thoracic features of an exemplar species. Osten-Sacken (1884) established a nomenclature of setae in Diptera, called chaetotaxy, and discussed certain setal arrangements in Asilidae. Later, Hardy (1930) provided a table summarizing the presence/absence of thoracic setae, including notopleural, postalar, supra-alar, and scutellar setae (see characters 88, 90–92, 94, 98), in Asilidae. Figure 62 provides an overview of thoracic setae coded in the present study.

Figure 61

Thorax of Philonicus albiceps (Asilinae) in lateral view with morphological features labeled. Wing and vestiture omitted; abbreviations see Materials and Methods. Scale line  =  1 mm.

Figure 62

Thorax of Philonicus albiceps (Asilinae) in lateral view with chaetotaxic characters mapped. Dotted lines indicate setose areas; wing and vestiture omitted; for abbreviations see Materials and Methods and variously numbered character descriptions. Scale line  =  1 mm.

68. Division of pronotum: (0) only a single sclerite present; (1) divided into ante- and postpronotum (length  =  1; CI  =  1.00; RI  =  1.00). In all Asilidae, the pronotum is divided into ante- and postpronotum (state 1; figs. 61, 64–65) (Martin, 1968). Among the outgroup taxa, the same arrangement is also found in Apioceridae, Mydidae, and Therevidae, whereas all remaining species possess a single sclerite (state 0).

Figure 63–65

Thoraces in anterodorsal view (anterior  =  bottom). 63. Nemomydas brachyrhynchus (Mydidae: Leptomydinae). 64. Dasyllis funebris (Laphriinae), dotted lines  =  acrostichal (medially) and dorsocentral (laterally) setae. 65. Lasiocnemus lugens (Leptogastrinae). Postpronotal lobes grey; vestiture omitted. Scale lines  =  1 mm.

69. Antepronotal setae: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  28; CI  =  0.07; RI  =  0.67). The antepronotum is usually setose in Asilidae and setae are absent only in Leptogastrinae except Acronyches maya and Nannodioctria sp. (formicaphaga) (Dioctriinae) (state 0). All Stichopogoninae; most Brachyrhopalinae, Trigonomiminae, Dioctriinae, and Laphriinae possess only regular setae (state 1) while macrosetae are developed in most species of the remaining taxa (state 2). Among the outgroup taxa, all character states are found.

70. Lateral postpronotal (pprn) setae: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  20; CI  =  0.10; RI  =  0.48). The lateral surfaces of the postpronotum are asetose in Leptogastrinae except Acronyches maya; Brachyrhopala ruficornis (Brachyrhopalinae); Nannodioctria sp. (formicaphaga) (Dioctriinae); and Goneccalypsis argenteoviridis and Loewinella nigripes (Laphriinae) (state 0). Most species possess at least some setae on the lateral postpronotum (state 1), but most Dasypogoninae, Stenopogoninae, and a few other species possess macrosetae (state 2). Among the outgroup taxa, Apioceridae and Rhaphiomidas maehleri (Mydidae) possess macrosetae and regular setae (state 2) whereas all remaining species possess only regular setae (state 1). All species in which the pronotum is not divided into an ante- and postpronotum (see character 68) are coded as inapplicable (-).

71. Dorsal fusion of postpronotal lobes with scutum: (0) entirely fused; (1) only partly fused, lobes elevated above scutum posteriorly (length  =  2; CI  =  0.50; RI  =  0.90). In all Asilidae the postpronotal lobes are entirely fused with the scutum although the sutures are still discernible (state 0; figs. 64–65). Among the outgroup taxa, the lobes are elevated above the scutal surface and not fused with it posteriorly in Bombyliidae, Mydidae, Nemestrinidae, and Scenopinidae (state 1; fig. 63).

72. Postpronotal lobes: (0) confined to lateral margin of scutum; (1) extending medially, but not anteriorly; (2) extending medially and anteriorly, nearly touching medially (length  =  3; CI  =  0.66; RI  =  0.95). Martin (1968) established the medially and anteriorly extending postpronotal lobes as an autapomorphy of his Leptogastridae. All Asilidae exhibit medially extending postpronotal lobes (state 1; fig. 64), but Leptogastrinae except Acronyches maya possess medially and anteriorly extending lobes (state 2; fig. 65). Among the outgroup taxa, the postpronotal lobes are usually confined to the lateral scutal margin (state 0) and only in the Scenopinidae do they also extend medially (state 1).

73. Dorsal postpronotal (pprn) setae: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  23; CI  =  0.08; RI  =  0.50). The postpronotal setae on the anterodorsal surface of the postpronotal lobes are usually present in Asilidae and only a few species lack regular setae (state 0). The majority of species possess setae (state 1) and most Dasypogoninae and Stenopogoninae possess macrosetae (state 2). Among the outgroup taxa, setae (state 1) and macrosetae (state 2) are present.

74. Lateral fusion of prosternum to proepisternum: (0) fused, broad prosternum; (1) fused, prosternum narrow above prothoracic coxa; (2) separated, sclerite triangular, pointed dorsally; (3) separated, sclerite square to rectangular, straight dorsally; (4) separated, prosternum square and proepisternum projecting medially (nearly touching prosternum) (length  =  31; CI  =  0.12; RI  =  0.76). Hardy (1930, 1948), Clements (1951), Hull (1962), Papavero (1973a), and Oldroyd (1974a) discussed in detail the development of the prosternum in relation to its lateral fusion to the proepisternum in Asilidae. Five character states are employed here and a completely fused, broad prosternum is present in all Stichopogoninae; Leptogastrinae, with the exception of Acronyches maya and Tipulogaster glabrata; and many Laphriinae (state 0; fig. 66). The prosternum and proepisternum are fused, but the prosternum is narrower above the coxa in all Trigonomiminae; some Asilinae and Laphriinae; and a few other species (state 1; fig. 69). A triangular, separate prosternum surrounded by membrane that is pointed dorsally is found in Tillobromatinae except Tillobroma punctipennis; Willistonininae except Willistonina bilineata; many Asilinae and Stenopogoninae; Afroholopogon peregrinus, Ceraturgus fasciatus, and Heteropogon manicatus (all Brachyrhopalinae); and Obelophorus terebratus (Phellinae) (state 2; fig. 68). A similar development is found in all Bathypogoninae and Dioctriinae; most Brachyrhopalinae and Dasypogoninae; a few Asilinae and Stenopogoninae; and some other species in that the prosternum is separate and surrounded by membrane but square to rectangular in shape with a straight dorsal margin (state 3; fig. 67). In Ommatius species (Ommatiinae) the prosternum is separate and square and the proepisternum is projecting medially to nearly touch the prosternum (state 4). Among the outgroup taxa, Apsilocephalidae, Mydidae, and Scenopinidae possess a fused prosternum and proepisternum (state 0) whereas they are fused and narrow above the coxa in Therevidae (state 1). The remaining outgroup taxa possess a free prosternum that is triangular in shape (state 2).

Figure 66–69

Thoraces in anterior view. 66. Stichopogon trifasciatus (Stichopogoninae). 67. Emphysomera pallidapex (Ommatiinae). 68. Pogonioefferia pogonias (Asilinae). 69. Pegesimallus laticornis (Dasypogoninae). Prosternum and proepisternum grey; vestiture omitted. Scale lines  =  1 mm.

75. Dorsal (anterior) margin of prosternum: (0) not elevated; (1) distinct flangelike projection (length  =  10; CI  =  0.10; RI  =  0.84). The dorsal (anterior) margin of the prosternum is usually not particularly pronounced (state 0; fig. 69), but in a few species a distinct dorsal flangelike projection is evident (state 1; fig. 66). It is found in all Laphriinae and Trigonomiminae; Leptogastrinae, with the exception of Acronyches maya and Tipulogaster glabrata; Stichopogoninae except Lasiopogon species; Dicolonus simplex and Dioctria species (Dioctriinae); Hypenetes stigmatias (Tillobromatinae); and Oligopogon sp. (Cape Recife) within Asilidae (state 1). Among the outgroup taxa, only Mydidae except Rhaphiomidas maehleri possess a flangelike projection (state 1).

76. Shape of propresternum: (0) arrowlike sclerite; (1) square sclerite; (2) triangular sclerite (length  =  3; CI  =  0.66; RI  =  0.90). Yeates (1994) used the shape of the propresternum in the phylogeny of Bombyliidae and emphasized the different shape within Asiloidea. All Asilidae possess a square propresternum that possesses two distinct sclerites bearing sensory setae (state 1; figs. 66–69). Among the outgroup taxa, an arrowlike sclerite is found in Apsilocephalidae, Bombyliidae, Nemestrinidae, and Therevidae (state 0). Apioceridae, Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri (Mydidae), and Scenopinidae possess a square sclerite similar to Asilidae (state 1) while the remaining Mydidae species possess a triangular sclerite (state 2).

77. Development of cervical sclerite: (0) not extending medially to cover sensory area on propresternum; (1) extending medially to cover sensory area on propresternum (length  =  2; CI  =  0.50; RI  =  0.87). The median margin of the cervical sclerite is closely associated with the propresternum, but usually does not come in contact with it (state 0; figs. 66–69). In Mydidae and Nemestrinidae, the median margin is extended and covers the sensory setae of the propresternum partly (state 1). In all Asilidae and all remaining outgroup taxa, the cervical sclerite is not touching the propresternum (state 0).

78. Anterior anepisternal setae: (0) absent; (1) erect; (2) erect to appressed, pointed anteriad (length  =  25; CI  =  0.08; RI  =  0.66). In Asilidae, the anterior and dorsal margins of the anepisternum possess usually short to long setae. These setae are absent in all Bathypogoninae and Dasypogoninae; most Stenopogoninae and Stichopogoninae; some Brachyrhopalinae and a few other species (state 0). In many Asilidae, the anterior anepisternal setae are erect and more or less perpendicular to the surface (state 1). In Leptogastrinae except Acronyches maya, and Dioctria species, Echthodopa pubera, and Myelaphus melas (all Dioctriinae), the setae are directed anteriorly and lightly appressed to the surface of the anepisternum (state 2). Other species of Dioctriinae lack these setae (state 0), i.e., Broticosia paramonovi, Eudioctria albius, and Nannodioctria sp. (formicaphaga), whereas Dicolonus simplex possesses only erect setae (state 1). Among the outgroup taxa, anepisternal setae are usually absent (state 0) and erect setae are present in Bombyliidae, Neorhaphiomidas sp. (Mandurah) (Mydidae), Nemestrinidae, and Phycus frommeri (Therevidae) (state 1).

79. Anterior proepimeral setae: (0) absent; (1) erect; (2) erect to appressed, pointed anteriad (length  =  11; CI  =  0.18; RI  =  0.64). The same development of anepisternal setae (see character 78) is also found on the proepimeron. Within Asilidae, only Pseudorus distendens (Dasypogoninae) and Lasiopogon species and Townsendia albomacula (Stichopogoninae) lack erect proepimeral setae (state 0). In Leptogastrinae except Acronyches maya, and Dioctria species, Echthodopa pubera, Eudioctria albius, and Myelaphus melas (all Dioctriinae), the proepimeral setae are lightly appressed and directed anteriorly (state 2). Among the outgroup taxa, proepimeral setae are erect in Apiocera aldrichii (Apioceridae), Apsilocephalidae, Bombylius major (Bombyliidae), Neorhaphiomidas sp. (Mandurah and Rhaphiomidas maehleri (Mydidae), and Nemestrinidae (state 1) and absent in the remaining taxa (state 0).

80. Macrosetae on superoposterior anepisternum: (0) absent; (1) present (length  =  5; CI  =  0.20; RI  =  0.81). Osten-Sacken (1884) indicated the presence of setae on the mesopleura in some Asilidae and Hermann (1920) found them in all Laphriinae he studied. Later, Papavero (1973a) used the presence of at least one macroseta on the superoposterior anepisternum as a character supporting the division of Laphriinae sensu previous authors into two taxa, i.e., Laphriinae and Laphystiinae. He pointed out that all Laphriinae possess this seta whereas it may be present or absent in his Laphystiinae. Macrosetae are present in all Andrenosomatini, Ctenotini, Laphriini, Atomosiini except Cerotainia albipilosa, Laxenecera albicincta, Nusa infumata, and Trichardis effrena (all Laphriinae); Michotamia aurata (Ommatiinae); and Acnephalum cylindricum (Willistonininae) (state 1). All remaining Asilidae and outgroup taxa lack macrosetae on the superoposterior anepisternum (state 0).

81. Setae on posterior basalare (post bas): (0) absent; (1) present (length  =  6; CI  =  0.16; RI  =  0.83). The posterior basalare is usually bare of setae in Asilidae (state 0). Setae are present in all Ommatiinae; Asilinae, with the exception of Dasophrys crenulatus, Eichoichemus pyrrhomystax, Lochmorhynchus albicans, Myaptex brachyptera, Neolophonotus bimaculatus, Nomomyia murina, Pogonioefferia pogonias, and Proctacanthus philadelphicus; as well as Lampria clavipes (Laphriinae) (state 1). Among the outgroup taxa, most species lack setae on the basalare (state 0) except Rhaphiomidas maehleri (Mydidae) (state 1).

82. Anatergal setae: (0) absent; (1) regular setae only (no macrosetae); (2) short, stout macrosetae (length  =  12; CI  =  0.16; RI  =  0.71). Papavero (1973a) used the presence/absence of anatergal setae (referred to as postscutellar or metanotal setae) to divide the Asilinae of earlier authors into Apocleinae and Asilinae, but later Artigas and Papavero (1997) treated these two taxa as a single group, highlighting the point that the distribution of anatergal setae is not constant within the taxa delimited. Anatergal setae are found here to be present in Asilinae, with the exception of Apoclea algira, Blepharotes splendidissimus, Dasophrys crenulatus, Eichoichemus pyrrhomystax, Lochmorhynchus albicans, Megaphorus pulchrus, Myaptex brachyptera, Neolophonotus bimaculatus, Nomomyia murina, Philodicus tenuipes, Pogonioefferia pogonias, Proctacanthus philadelphicus, and Promachus amastrus; as well as Megapoda labiata and Pseudorus distendens (Dasypogoninae); Dicolonus simplex (Dioctriinae); Aphestia annulipes, Dasyllis funebris, Laphria aktis, and Pilica formidolosa (all Laphriinae); and Daspletis stenoura and Microstylum sp. (Karkloof) (Stenopogoninae) (state 1). Hermann (1912) highlighted the short, stout macrosetae on the anatergite of Atomosiini (Laphriinae) as a characteristic feature and these setae are found here to be present in Atomosiini with the exception of Aphestia annulipes, Orthogonis scapularis, and Smeryngolaphria numitor (state 2). Among the outgroup taxa, most Mydidae possess anatergal setae as well (state 1), but the remaining taxa possess asetose anatergites (state 0).

83. Katatergal setae: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  19; CI  =  0.10; RI  =  0.62). The katatergite is usually setose with a mixture of regular setae and macrosetae in Asilidae (state 2). Katatergal setae are only absent in Gonioscelis ventralis and Stenopogon species (Stenopogoninae) and regular setae are present in Leptogastrinae except Lasiocnemus lugens; as well as Megaphorus pulchrus (Asilinae); Cyrtophrys attenuatus, Deromyia fuscipennis, Megapoda labiata, Molobratia teutonus, Pegesimallus laticornis, Pseudorus distendens, and Senobasis corsair (all Dasypogoninae); Atomosia puella, Cerotainia albipilosa, and Hyperechia nigripennis (all Laphriinae); Phellus olgae (Phellinae); Scleropogon subulatus (Stenopogoninae); and some Brachyrhopalinae (state 1). Among the outgroup taxa, the katatergite is usually asetose (state 0) and Bombylius major (Bombyliidae), Afroleptomydas sp. (Clanwilliam), Nemomydas brachyrhynchus, Neorhaphiomidas sp. (Mandurah), and Pseudonomoneura hirta (all Mydidae), Nemestrinidae, and Hemigephyra atra (Therevidae) possess katatergal setae (state 1). Phycus frommeri (Therevidae) possesses katatergal setae and macrosetae (state 2).

84. Katatergite shape: (0) slightly elevated sclerite; (1) cone-shaped sclerite (length  =  2; CI  =  0.50; RI  =  0.00). The katatergite is usually a slightly elevated sclerite dorsal of the metathoracic spiracle (state 0), but in Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri (Mydidae) it is a pronounced, cone-shaped sclerite (state 1).

85. Setae on posterior meron + metanepisternum: (0) absent; (1) regular setae only (no macrosetae); (2) at least one macroseta among setae (length  =  21; CI  =  0.09; RI  =  0.77). The posterior meron + metanepisternum, often referred to as the metapleura, is usually asetose in Asilidae (state 0), but all Asilinae, Laphriinae, and Ommatiinae possess at least regular setae (state 1). Papavero (1973a) stated that these setae are present only in his Apocleinae, Asilinae, and Ommatiinae. Macrosetae among the setae are present in many Asilinae and Ommatiinae and Daspletis stenoura (Stenopogoninae) (state 2). Other Asilidae possessing regular setae are Heteropogon manicatus and Holopogon priscus (Brachyrhopalinae); Megapoda labiata and Pseudorus distendens (Dasypogoninae); Dicolonus simplex and Echthodopa pubera (Dioctriinae); Phellus olgae (Phellinae); Creolestes nigribarbis, Microstylum sp. (Karkloof), and Prolepsis tristis (all Stenopogoninae); Lycostommyia albifacies (Tillobromatinae); and Damalis monochaetes and Damalis sp. (Palatupana) (Trigonomiminae) (state 1). Among the outgroup taxa, only Bombyliidae and Nemestrinidae possess regular setae (state 1).

86. Setae on superoposterior anepimeron: (0) absent; (1) regular setae only (no macrosetae); (2) at least one macroseta among setae (length  =  22; CI  =  0.09; RI  =  0.70). Similar to the setose meron + metanepisternum, the anepimeron can possess setae on its superoposterior surface as well. These setae are found in all Ommatiinae; Asilinae except Apoclea algira; Megapoda labiata and Pseudorus distendens (Dasypogoninae); most Laphriinae; Lasiocnemus lugens (Leptogastrinae); Daspletis stenoura and Microstylum sp. (Karkloof) (Stenopogoninae); Phellus olgae (Phellinae); and Lycostommyia albifacies (Tillobromatinae) (state 1). Asilus crabroniformis, Myaptex brachyptera, Neoitamus cyanurus, Satanas gigas, and Threnia carbonaria (all Asilinae), and Emphysomera species, Michotamia aurata, and Afroestricus chiastoneurus (Ommatiinae) possess macrosetae and regular setae (state 2). Among the outgroup taxa, setae are found in Bombylius major (Bombyliidae), Mydidae, with the exception of Nemomydas brachyrhynchus, Neorhaphiomidas sp. (Mandurah), and Rhaphiomidas maehleri, and Nemestrinidae (state 1).

87. Short, stout, erect macrosetae on anterior mesonotum: (0) absent; (1) present (length  =  1; CI  =  1.00; RI  =  1.00). Londt (1988) used the short, stout, erect macrosetae on the anterior mesonotum as distinguishing characteristics for some Afrotropical Atomosiini genera. These setae are present only in the Afrotropical Goneccalypsis argenteoviridis and Loewinella nigripes (Laphriinae) (state 1). Among the outgroup taxa, all species lack these setae (state 0).

88. Notopleural (npl) setae: (0) absent; (1) one seta; (2) two setae; (3) three or more setae; (4) single, short stout macroseta (length  =  42; CI  =  0.09; RI  =  0.60). Notopleural setae are usually present in Asilidae (states 1, 2, and 3), but absent in Trigonomiminae except Damalis sp. (Palatupana) (state 0). The number of setae varies considerably and states (1), (2), and (3) are found widely within Asilidae. Chrysopogon pilosifacies and Codula limbipennis (Brachyrhopalinae) possess particularly short, stout macrosetae that are coded in a separate state (state 4). The following species are coded as polymorphic as more than one state is observed in different individuals, i.e., states (0) and (1) in Myelaphus melas (Dioctriinae), states (1) and (2) in Protometer sp. (El Tuparro) (Laphriinae), states (2) and (3) in Laxenecera albicincta (Laphriinae); Scylaticus costalis (Stenopogoninae); Lasiopogon cinctus (Stichopogoninae); and Lycostommyia albifacies (Tillobromatinae). Among the outgroup taxa, Apsilocephalidae and Hemigephyra atra (Therevidae) possess a single notopleural (npl) seta (state 1), Phycus frommeri (Therevidae) possesses two npl setae (state 2), and Apioceridae and Rhaphiomidas maehleri (Mydidae) possess three or more npl setae (state 3). All remaining species lack any npl setae (state 0).

89. Transverse (trn) suture: (0) entirely absent (not discernible); (1) distinct, at least on lateral surface (length  =  1; CI  =  1.00; RI  =  1.00). Yeates (1994) highlighted the absence of the transverse suture in the genus Neorhaphiomidas Norris, 1936 (Mydidae) and Acroceridae. The transverse suture is usually well developed and present in all Asilidae (state 1). In the Mydidae except Rhaphiomidas maehleri, it is absent or not discernible (state 0) (fig. 63).

90. Supra-alar (spa) setae: (0) absent; (1) present (length  =  9; CI  =  0.11; RI  =  0.68). Supra-alar setae are usually present in Asilidae (state 1) and are absent only in all Trigonomiminae; Chrysopogon pilosifacies, Codula limbipennis, Dasycyrton gibbosus, Holopogon priscus, Leptarthrus brevirostris (all Brachyrhopalinae); Pseudorus distendens (Dasypogoninae); and Lissoteles aquilonius and Stichopogon trifasciatus (Stichopogoninae) (state 0). Among the outgroup taxa, supra-alar setae are absent in Bombyliidae, Mydidae except Rhaphiomidas maehleri, Nemestrinidae, and Scenopinidae (state 0).

91. Postalar (pal) setae: (0) absent; (1) present (length  =  12; CI  =  0.08; RI  =  0.66). Postalar setae are usually present in Asilidae (state 1) and are only absent in Afroholopogon peregrinus, Chrysopogon pilosifacies, Codula limbipennis, Dasycyrton gibbosus, Leptarthrus brevirostris (all Brachyrhopalinae); Pseudorus distendens (Dasypogoninae); Dioctriinae, with the exception of Broticosia paramonovi, Dicolonus simplex, and Echthodopa pubera; Hyperechia nigripennis (Laphriinae); Leptogastrinae, with the exception of Acronyches maya and Leptogaster cylindrica; and Trigonomiminae except Damalis annulata (state 0). Among the outgroup taxa, postalar setae are absent in Bombyliidae, Mydidae except Rhaphiomidas maehleri, and Nemestrinidae (state 0).

92. Presutural dorsocentral (dc) setae: (0) absent; (1) present (length  =  43; CI  =  0.02; RI  =  0.44). The development of pre- and postsutural dorsocentral setae can be informative for identification of robber flies and has been reviewed by Hardy (1930). Both character states are widely found within Asilidae. It is sometimes difficult to discern the presutural dorsocentral setae and also ambiguous as to how long the seta should be to call it a well-developed dorsocentral seta. Among the outgroup taxa, presutural dorsocentral setae are present only in Apsilocephalidae and Mydidae, with the exception of Mydas clavatus, Nemomydas brachyrhynchus, and Rhaphiomidas maehleri (state 1).

93. Presutural acrostichal (acr) setae (in regular rows): (0) absent; (1) present (length  =  12; CI  =  0.08; RI  =  0.15). Presutural acrostichal setae arranged in regular rows are almost always present in Asilidae (state 1) and only absent in Deromyia fuscipennis, Megapoda labiata, and Senobasis corsair (all Dasypogoninae); Goneccalypsis argenteoviridis (Laphriinae); Tipulogaster glabrata (Leptogastrinae); Ommatius incurvatus (Ommatiinae); and Stichopogon elegantulus (Stichopogoninae) (state 0). Among the outgroup taxa, only a few species lack presutural acrostichal setae (state 0).

94. Postsutural dorsocentral (dc) setae: (0) absent; (1) present (length  =  16; CI  =  0.06; RI  =  0.31). Postsutural dorsocentral setae are usually developed in Asilidae (state 1) and absent only in Codula limbipennis (Brachyrhopalinae); Molobratia teutonus and Pegesimallus laticornis (Dasypogoninae); Dioctria species and Myelaphus melas (Dioctriinae); Hoplistomerus nobilis, Lamyra gulo, Perasis transvaalensis, Protometer sp. (El Tuparro), and Trichardis effrena (all Laphriinae); Tipulogaster glabrata (Leptogastrinae); Stichopogon trifasciatus (Stichopogoninae); and Holcocephala species (Trigonomiminae) (state 0). Among the outgroup taxa, most species possess postsutural dorsocentral setae (state 1).

95. Sockets surrounding scutal setae: (0) absent (setae without large sockets); (1) present (setae with large sockets) (length  =  6; CI  =  0.16; RI  =  0.58). The base of the scutal setae and macrosetae is usually not differentiated and no particular indentation is evident (state 0). Hermann (1912) pointed out that many Atomosiini possess large setal sockets on mesonotum and abdominal tergites (see character 151). In Leptarthrus brevirostris (Brachyrhopalinae); Dioctria species (Dioctriinae); Atomosiini, with the exception of Orthogonis scapularis and Smeryngolaphria numitor, Hoplistomerus nobilis, Stiphrolamyra angularis, and Trichardis effrena (all Laphriinae); and Euscelidia pulchra (Leptogastrinae), large sockets are evident (state 1). Among the outgroup taxa, setal sockets are not developed (state 0).

96. Development of medial setae on posterior scutum (between dc setae): (0) absent; (1) present, directed posteriorly; (2) present, directed anteriorly; (3) present, appressed to surface, directed posteriorly (length  =  21; CI  =  0.14; RI  =  0.43). The posterior part of the mesonotum possesses generally additional setae between the two rows of dorsocentral setae. These setae are usually oriented posteriorly in Asilidae (state 1). The setae are absent in Cyrtophrys attenuatus, Dasypogon diadema, and Megapoda labiata (all Dasypogoninae); Ablautus coquilletti (Willistonininae); and Coleomyia setigera (state 0). In Leptogastrinae except Beameromyia bifida; Apoclea algira (Asilinae); Goneccalypsis argenteoviridis and Zabrops tagax (Laphriinae); Lissoteles aquilonius and Stichopogon species (Stichopogoninae); Hypenetes stigmatias (Tillobromatinae); Holcocephala species and Rhipidocephala sp. (Harold Johnson) (Trigonomiminae); and Acnephalum cylindricum (Willistonininae) these setae are oriented anteriorly (state 2). In Choerades bella, Lampria clavipes, Laphria aktis, and Maira aenea (Laphriinae), these setae are oriented posteriorly and more or less tightly appressed to the thorax (state 3). Among the outgroup taxa, the setae are absent in Scenopinidae (state 0) and directed anteriorly in Mydidae, with the exception of Mitrodetus dentitarsis and Rhaphiomidas maehleri, and Nemestrinidae (state 2). All remaining outgroup taxa exhibit state 1.

97. Size of mesonotal scutellum: (0) small, mesopostnotum visible in dorsal view; (1) large, mesopostnotum not visible in dorsal view (length  =  4; CI  =  0.25; RI  =  0.80). Yeates and Irwin (1996) established that the small scutellum is an autapomorphy of a subordinate taxon of Mydidae. In Asilidae, the scutellum is generally large and the mesopostnotum is not visible in dorsal view (state 1). In Cabasa pulchella (Brachyrhopalinae); Blepharepium cajennensis, Cyrtophrys attenuatus, Deromyia fuscipennis, Megapoda labiata, Molobratia teutonus, Pegesimallus laticornis, Pseudorus distendens, and Senobasis corsair (all Dasypogoninae); and Plesiomma sp. (Guanacaste) (Stenopogoninae), the scutellum is small, so that the mesopostnotum is visible in dorsal view (state 0). Among the outgroup taxa, a small scutellum is found in Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri (state 0).

98. Apical scutellar (ap sctl) setae: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  31; CI  =  0.06; RI  =  0.58). Hardy (1930) (within Asilidae) and Yeates and Irwin (1996) (within Apioceridae and Mydidae) examined the apical scutellar setae. In Asilidae, there are usually some setae present, e.g., Dioctriinae, Leptogastrinae, Trigonomiminae, and many Brachyrhopalinae (state 1), but the majority of robber flies possess macrosetae as well as regular setae on the apical margin of the scutellum (state 2). No apical scutellar setae are developed in Blepharepium cajennensis, Cyrtophrys attenuatus, Deromyia fuscipennis, Molobratia teutonus, and Senobasis corsair (all Dasypogoninae); Eudioctria albius and Myelaphus melas (Dioctriinae); Perasis transvaalensis, Protometer sp. (El Tuparro), and Trichardis effrena (all Laphriinae); Michotamia aurata (Ommatiinae); Plesiomma sp. (Guanacaste) (Stenopogoninae); and Holcocephala species, Rhipidocephala sp. (Harold Johnson), and Trigonomima sp. (anamaliensis) (all Trigonomiminae) (state 0). Among the outgroup taxa, Bombyliidae, Nemestrinidae, and Scenopinidae possess apical scutellar setae (state 1) and Apioceridae, Apsilocephalidae, and Therevidae possess macrosetae as well as setae (state 2). Within Mydidae, Rhaphiomidas maehleri possesses macrosetae and regular setae (state 2), Mitrodetus dentitarsis and Neorhaphiomidas sp. (Mandurah) possess only regular setae (state 1) and in the remaining species apical scutellar setae are absent (state 0).

99. Discal scutellar (ds sctl) setae: (0) absent; (1) present (length  =  18; CI  =  0.05; RI  =  0.65). The scutellar disc is generally setose in Asilidae (state 1), but a few taxa lack any setae on the scutellum. These are all Willistonininae; Dasypogoninae, with the exception of Megapoda labiata, Pegesimallus laticornis, and Pseudorus distendens; Stenopogoninae except Enigmomorphini; Stichopogoninae, with the exception of Lissoteles aquilonius and Stichopogon punctum; and a few other species of Brachyrhopalinae; Dioctriinae; Leptogastrinae; and Trigonomiminae (state 0). Among the outgroup taxa, only Afroleptomydas sp. (Clanwilliam), Nemomydas brachyrhynchus, and Pseudonomoneura hirta (all Mydidae), Hemigephyra atra (Therevidae), and Scenopinidae lack discal scutellar setae (state 0).

100. Coxae and metkatepisternum (in ventral view): (0) metkatepisternum small and not visible between mesothoracic and metathoracic coxae; (1) metkatepisternum large and visible between mesothoracic and metathoracic coxae (length  =  7; CI  =  0.14; RI  =  0.81). Yeates and Irwin (1996) postulated the large metkatepisternum (referred to as metepisternum) as an autapomorphy for a subordinate taxon of Mydidae. In Asilidae, the metkatepisternum is usually small and not visible between the mes- and metathoracic coxae (state 0). In Laphriinae, with the exception of Aphestia annulipes, Atomosia puella, Cerotainia albipilosa, Goneccalypsis argenteoviridis, Loewinella nigripes, and Psilocurus modestus; as well as Megapoda labiata and Pseudorus distendens (Dasypogoninae), and Acronyches maya (Leptogastrinae), the metkatepisternum is large and visible in ventral view between the mes- and metathoracic coxae (state 1). Among the outgroup taxa, Mydidae except Rhaphiomidas maehleri and Nemestrinidae also possess a large metkatepisternum (state 1).

101. Metkatepisternal setae: (0) absent; (1) present (length  =  9; CI  =  0.11; RI  =  0.85). The metkatepisternum is usually asetose in Asilidae (state 0), but all Asilinae and Ommatiinae as well as Broticosia paramonovi (Dioctriinae); Andrenosomatini except Andrenosoma cyrtoxys, Aphestia annulipes, and Lamyra gulo (all Laphriinae); and Damalis monochaetes (Trigonomiminae), possess setae on this sclerite (state 1). Among the outgroup taxa, all species that possess a large metkatepisternum (see character 100) also possess setae (state 1).

102. Postmetacoxal bridge: (0) absent, postmetacoxal area entirely membranous; (1) partly present laterally, membranous area medially; (2) present, visible suture between lateral halves; (3) present, entirely sclerotized (length  =  26; CI  =  0.11; RI  =  0.56). Several authors (e.g., Hull, 1962; Papavero, 1973a) have emphasized the development of sclerotization in the postmetacoxal area in Asilidae. It is here coded into three different states. Most Asilidae possess an entirely membranous postmetacoxal area with a straight lateral margin (state 0). A partly developed postmetacoxal bridge in which the lateral sclerites fold medially, but where the median area is still membranous to a large extent, is found in Austrosaropogon nigrinus (Brachyrhopalinae); Pseudorus distendens (Dasypogoninae); Leptogastrinae, with the exception of Acronyches maya, Euscelidia pulchra, and Lasiocnemus lugens; as well as Lissoteles aquilonius and Stichopogon species (Stichopogoninae); Damalis annulata and Damalis monochaetes (Trigonomiminae); most Laphriinae; and a number of Asilinae (state 1). The postmetacoxal bridge is present, but a median suture is still visible dividing the two halves in Euscelidia pulchra and Lasiocnemus lugens (Leptogastrinae) and Townsendia albomacula (Stichopogoninae) (state 2). An entirely complete postmetacoxal bridge is developed in Atomosiini, with the exception of Orthogonis scapularis and Smeryngolaphria numitor (Laphriinae); Ommatiinae; Rhabdogaster pedion (Brachyrhopalinae); and Damalis sp. (Palatupana) (Trigonomiminae) (state 3). Among the outgroup taxa, all species possess an entirely membranous postmetacoxal area (state 0).

103. Shape of setae on legs: (0) only circular setae; (1) dorsoventrally flattened setae and circular setae (length  =  3; CI  =  0.33; RI  =  0.50). Setae on all legs are usually circular in cross section in Asilidae (state 0), but Willistonininae except Willistonina bilineata possess dorsoventrally flattened setae as well as circular setae on the legs (state 1). Among the outgroup taxa, only Bombylius major (Bombyliidae) possesses dorsoventrally flattened setae (state 1).

104. Lateral depression on prothoracic coxa: (0) absent; (1) present (length  =  3; CI  =  0.33; RI  =  0.85). The lateral surface of the prothoracic coxae is usually markedly depressed in Asilidae so that the coxae are not circular in cross section (state 1). This depression appears to be necessary to fold the prothoracic femora tightly against the coxae and thorax when perching. Only in Leptogastrinae do the coxae appear to be circular in cross section (state 0). Among the outgroup taxa, Apioceridae and Mydidae except Pseudonomoneura hirta possess indented coxae as well (state 1) whereas all remaining taxa possess circular coxae (state 0).

105. Setae on anteroventral prothoracic tibiae: (0) absent; (1) at least three setae present (length  =  10; CI  =  0.10; RI  =  0.72). The anteroventral surface of the prothoracic tibiae in Asilidae are found to possess at least three setae facing ventromedially. These setae are absent only in Afroholopogon peregrinus and Heteropogon manicatus (Brachyrhopalinae); Lissoteles aquilonius and Stichopogon species (Stichopogoninae); Stenopogoninae, with the exception of Ancylorhynchus fulvicollis, Connomyia varipennis, Daspletis stenoura, Gonioscelis ventralis, Scylaticus costalis, and Stenopogon species; and Willistonininae except Sisyrnodytes sp. (Gamka) (state 0). Among the outgroup taxa, the setae are absent in all species (state 0).

106. Projections on prothoracic tibiae: (0) absent; (1) small S-shaped spur, originating posteroventrally; (2) large spine, originating anteroventrally (length  =  8; CI  =  0.25; RI  =  0.70). One of the most obvious characters used in classifications on Asilidae has been the development of spurs or spines on the prothoracic tibiae (e.g., Hermann 1912, 1920; Hull, 1962; Papavero, 1973a; Oldroyd, 1974a) and Hermann (1912) based an entire taxon on the presence of projections, i.e., Acanthocneminae (unavailable name: Sabrosky, 1999; Dikow, 2004). The majority of Asilidae species lack any projections on the prothoracic tibiae with the exception of straight macrosetae or setae. The large spine found in Molobratia teutonus (fig. 70) is easy to observe, but the smaller, S-shaped spur in Leptarthrus brevirostris (fig. 72) is difficult to see and to distinguish from the surrounding macrosetae. A detailed examination of the different tibial projections reveals that the large spines are always originating on the posteroventral surface whereas the small, S-shaped spurs originate on the anteroventral surface of the prothoracic tibiae. Hull (1962) and Oldroyd (1974a) recognized the different shapes of the projections, but not their different origin on the prothoracic tibiae. Both projections are considered here to be nonhomologous and coded as two different states. A small, S-shaped spur is found in Cophura brevicornis, Leptarthrus brevirostris, and Nicocles politus (all Brachyrhopalinae) and Thereutria amaraca (Dasypogoninae) (state 1). A large spine is found in Dasypogoninae, with the exception of Archilestris magnificus and Thereutria amaraca; as well as Austrosaropogon nigrinus, Brachyrhopala ruficornis, Cabasa pulchella, and Chrysopogon pilosifacies (all Brachyrhopalinae) (state 2). Among the outgroup taxa, tibial projections are absent (state 0).

Figure 70–72

SEM micrographs of prothoracic tibiae in lateral view. 70. Molobratia teutonus (Dasypogoninae). 71. Molobratia teutonus, detail of spine with spicules on tarsomere. 72. Leptarthrus brevirostris (Brachyrhopalinae). Scale lines  =  200 µm.

107. Base of spine/spur on prothoracic tibia: (0) spine/spur originating directly from tibia; (1) spine/spur originating from projection of tibia (length  =  3; CI  =  0.33; RI  =  0.77). The base of the prothoracic tibial projection varies in species that possess these projections and Hull (1962) mentioned the pronounced and enlarged base in Molobratia. The spine/spur originates directly from the tibia in all Brachyrhopalinae and Blepharepium cajennensis, Diogmites grossus, Lestomyia fraudiger, Neodiogmites melanogaster, and Thereutria amaraca (all Dasypogoninae) (state 0). In the remaining Dasypogoninae, the spine originates from a prominent projection on the tibia (state 1; figs. 70–71). All species not possessing spines/spurs are coded as inapplicable (-) (see character 106).

108. Proximal prothoracic tarsomere: (0) cylindrical; (1) cylindrical with short spicules; (2) expanded proximoventrally with stout spicules (length  =  7; CI  =  0.28; RI  =  0.70). The proximal prothoracic tarsomere is usually simple and cylindrical without any projections in Asilidae (state 0). Hull (1962) commented on the different development of the prothoracic tarsomere in species possessing tibial projections. In Dasypogoninae except Archilestris magnificus and a few species of Brachyrhopalinae, which possess a large spine on the tibia, the proximal tarsomere often possesses projections in forms of spicules or expanded protuberances. In Diogmites grossus, Lestomyia fraudiger, Neodiogmites melanogaster, and Thereutria amaraca (all Dasypogoninae) and Austrosaropogon nigrinus, Brachyrhopala ruficornis, Cabasa pulchella, and Chrysopogon pilosifacies (all Brachyrhopalinae) the proximal tarsomere is cylindrical, but bears short spicules (state 1; fig. 71). In the remaining Dasypogoninae, the tarsomere is expanded proximoventrally and stout spicules are positioned on this elevated protuberance (state 2). Among the outgroup taxa, all species possess a cylindrical proximal tarsomere (state 0).

109. Orientation of prothoracic and mesothoracic coxae: (0) distinctly directed posteriorly; (1) more or less directed ventrally (sometimes even anteriorly) (length  =  7; CI  =  0.14; RI  =  0.87). In Asilidae, the pro- and mesothoracic coxae are usually directed more or less ventrally (sometimes even anteriorly), which permits them to catch and hold prey directly under the thorax (state 1; fig. 61). In Ceraturgus fasciatus (Brachyrhopalinae); Laphriinae except Atomosiini sensu stricto; Megapoda labiata and Pseudorus distendens (Dasypogoninae); Creolestes nigribarbis (Stenopogoninae); Lissoteles aquilonius and Stichopogon species (Stichopogoninae); and Tillobroma punctipennis (Tillobromatinae), the coxae are directed posteriorly (state 0). Among the outgroup taxa, the coxae are also directed posteriorly (state 0).

110. Macrosetae on antero-dorsal surface of mesothoracic tibiae: (0) absent; (1) one seta present; (2) several setae in one row present (length  =  18; CI  =  0.11; RI  =  0.36). Hennig (1973) postulated the densely bristled legs as an autapomorphy of Asilidae. I, however, am unable to find any particular setal arrangement or characters of the legs that could be easily homologized for the entire taxon. The row of macrosetae on the anterodorsal surface of the mesothoracic tibia is the only exception. Most Asilidae possess several setae at this position (state 2), but all Leptogastrinae; Blepharotes splendidissimus, Colepia rufiventris, Heligmonevra laevis, Megaphorus pulchrus, Pogonioefferia pogonias, and Zosteria rosevillensis (all Asilinae); Myelaphus melas (Dioctriinae); and Laphria species (Laphriinae) lack any strong macrosetae on the mesothoracic tibia (state 0). Apoclea algira, Lochmorhynchus albicans, Nomomyia murina, and Philodicus tenuipes (all Asilinae) possess only a single seta (state 1). Among the outgroup taxa, most species possess a few setae (state 2) and only Nemomydas brachyrhynchus and Neorhaphiomidas sp. (Mandurah) (Mydidae), Nemestrinidae, and Scenopinidae lack setae (state 0). Asilus sericeus and Lycomya germainii (Asilinae, 0/1) and Ommatius tibialis (Ommatiinae, 1/2) are coded as polymorphic as several individuals possess different numbers of setae.

111. Orientation of metathoracic coxae: (0) distinctly directed ventrally to posteriorly; (1) directed anteriorly (length  =  1; CI  =  1.00; RI  =  1.00). The metathoracic coxae are directed ventrally to posteriorly in most Asilidae and all outgroup taxa (state 0). In Leptogastrinae except Acronyches maya, the coxae are distinctly directed anteriorly (state 1). Among the outgroup taxa, all species exhibit state 0.

112. Setae on lateral surface of metathoracic coxa: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  22; CI  =  0.09; RI  =  0.70). Robber flies usually possess setae on the lateral surface of the metathoracic coxae and often these setae are developed as strong macrosetae. Setae are absent in Tipulogaster glabrata and Euscelidia pulchra (Leptogastrinae); Townsendia albomacula (Stichopogoninae); and Coleomyia setigera (state 0). Macrosetae are generally found within Asilinae; Dasypogoninae; Ommatiinae; and Stenopogoninae (state 2), whereas the remaining taxa possess generally only regular setae (state 1). Among the outgroup taxa, all three character states are found.

113. Protuberance on anterior surface of metathoracic coxa: (0) absent; (1) present, protuberance blunt distally; (2) present, peglike protuberance pointed distally (length  =  16; CI  =  0.12; RI  =  0.39). Cannings (2002) observed the blunt protuberance on the anterior surface of the metathoracic coxa of some Stichopogoninae in contrast to species of Lasiopogon, and Londt (2005a) based the description of the Afrotropical genus Ischiolobos 82Londt, 2005 (placed in Stenopogoninae sensu Papavero, 1973a), on the presence of a blunt protuberance on the metathoracic coxa. The majority of Asilidae do not possess any protuberance on the metathoracic coxae (state 0). In Bathypogoninae; as well as Cophura brevicornis, Holopogon priscus, Metapogon punctipennis, Nannocyrtopogon lestomyiformis (all Brachyrhopalinae); Creolestes nigribarbis (Stenopogoninae); Stichopogon species (Stichopogoninae); Hypenetes stigmatias and Tillobroma punctipennis (Tillobromatinae); Trichoura sp. (Tierberg) and Willistonina bilineata (Willistonininae); and Coleomyia setigera, a blunt protuberance is present (state 1). In Dioctria species, Eudioctria albius, and Nannodioctria sp. (formicaphaga) (all Dioctriinae), a peglike, pointed protuberance is developed (state 2). Among the outgroup taxa, all Apioceridae, Apsilocephalidae, Scenopinidae, and Therevidae possess a blunt protuberance on the metathoracic coxae (state 1).

114. Setae on median surface of metathoracic trochanter: (0) absent; (1) regular setae only (no macrosetae); (2) macrosetae and regular setae (length  =  28; CI  =  0.07; RI  =  0.64). All Asilidae species possess at least regular setae on the median surface of the metathoracic trochanter (state 1) and many species of Asilinae; Bathypogoninae; Dioctriinae; Ommatiinae; Stenopogoninae; Stichopogoninae; and Tillobromatinae possess macrosetae (state 2). Among the outgroup taxa, these setae are absent in Rhaphiomidas maehleri (Mydidae) and most other species only possess setae (state 1).

115. Shape of median surface of metathoracic trochanter: (0) cylindrical, medially without expansion; (1) medially with posteriorly directed expansion (length  =  5; CI  =  0.20; RI  =  0.33). The metathoracic trochanter is usually cylindrical in Asilidae (state 0), but in Lampria clavipes, Laxenecera albicincta, and Maira aenea (all Laphriinae), as well as Damalis species (Trigonomiminae), the medial surface of the trochanter is expanded (state 1). Among the outgroup taxa, only Mydas clavatus (Mydidae) possesses a medially expanded trochanter (state 1).

116. Setation on ventral surface of metathoracic femur: (0) absent or only regular setae present; (1) macrosetae present; (2) short, stout, cone-shaped macrosetae present (length  =  30; CI  =  0.06; RI  =  0.66). Woodley (1989) postulated that the short, stout, cone-shaped macrosetae on the ventral surface of the metathoracic femur found in many Mydidae is an autapomorphy of this taxon. Although Asilidae usually possess setae on the ventral surface of the metathoracic femur (state 0), only within Asilinae; Ommatiinae; and Stenopogoninae do the majority of species possess macrosetae (state 1). Brachyrhopala ruficornis (Brachyrhopalinae); Hoplistomerus nobilis, Lampria clavipes, and Trichardis effrena (all Laphriinae); Hypenetes stigmatias and Lycostommyia albifacies (Tillobromatinae); and Damalis monochaetes and Damalis sp. (Palatupana) (Trigonomiminae) possess short, stout, cone-shaped macrosetae (state 2). Among the outgroup taxa, cone-shaped macrosetae are only found in Afroleptomydas sp. (Clanwilliam) and Mydas clavatus within Mydidae (state 2) although all remaining species of Mydidae and Apioceridae possess macrosetae on the ventral metathoracic femora (state 1).

117. Cuticular facets medially on proximal surface of metathoracic femur: (0) absent; (1) present (length  =  1; CI  =  1.00; RI  =  1.00). Clements (1985) observed the presence of cuticular facets on the metathoracic femora in all known species of Codula and most species of Chrysopogon and absence of these in Chryseutria Hardy, 1928, and Opseostlengis White, 1914, all belonging to his concept of Chrysopogonini. These cuticular facets are found here to be present only in Chrysopogon pilosifacies and Codula limbipennis (state 1). Among the outgroup taxa, cuticular facets are lacking (state 0).

118. Fringe of parallel setae ventrally on metathoracic femora and tibia: (0) absent; (1) present on tibiae only; (2) present on femora and tibiae (length  =  4; CI  =  0.50; RI  =  0.83). Fisher (personal commun.) pointed out the fringe of parallel setae on the ventral metathoracic femora and tibiae in many Dioctriinae and Atomosiini (Laphriinae). These setae are present on the metathoracic tibiae only in Aphestia annulipes (Laphriinae) (state 1) and present on femora and tibiae in Dioctriinae, with the exception of Broticosia paramonovi and Myelaphus melas; and the remaining Atomosiini (all Laphriinae) (state 2). The fringe of parallel setae are absent in all remaining Asilidae and all outgroup taxa (state 0).

119. Shape of metathoracic tibia: (0) straight; (1) arched laterally; (2) arched medially (length  =  10; CI  =  0.20; RI  =  0.50). The metathoracic tibia is usually straight in Asilidae (state 0), but in a number of species of Dasypogoninae and Laphriinae, it is distinctly arched laterally or medially. Laterally arched tibiae are found in Andrenosoma cyrtoxys, Dasyllis funebris, Hoplistomerus nobilis, Hyperechia nigripennis, Lampria clavipes, Laphria species, Laphystia aegyptiaca, Laxenecera albicincta, Maira aenea, Nusa infumata, Pilica formidolosa, and Trichardis effrena (all Laphriinae) (state 1), and medially arched tibiae are found in Blepharepium cajennensis, Cyrtophrys attenuatus, Deromyia fuscipennis, and Senobasis corsair (all Dasypogoninae) (state 2). Among the outgroup taxa, Mydas clavatus (Mydidae) is the only species possessing a laterally arched tibia (state 1).

120. Length of proximal prothoracic tarsomere: (0) shorter or same length as following two tarsomeres combined; (1) longer than following two tarsomeres combined (length  =  24; CI  =  0.04; RI  =  0.70). The length of the proximal tarsomeres relative to the following tarsomeres has been used as a diagnostic and phylogenetic character within Mydidae (e.g., Papavero and Wilcox, 1974; Yeates and Irwin, 1996), but has not been evaluated across Asilidae. Here, it is only determined whether the proximal tarsomere is shorter or longer than the two following tarsomeres combined, to avoid morphometric calculations that cannot be coded as discrete character states. See table 3 for character state distribution within Asilidae. Among the outgroup taxa, most species possess longer proximal tarsomeres (state 1) and only Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah), Opomydas townsendi, and Rhaphiomidas maehleri possess shorter proximal tarsomeres (state 0).

Table 3

Distribution of lengths of proximal tarsomeres (chars. 120–122) Note that a species exhibiting state (0) on the prothoracic leg might not exhibit the same state on another leg, for example, and the columns are just counts of species; see character matrix in appendix 1 for details; number in parentheses following the taxon name equals number of species of the taxon included in the present study; state 0  =  shorter or equal to the length of two following tarsomeres combined; state 1  =  longer than two following tarsomeres combined

121. Length of proximal mesothoracic tarsomere: (0) shorter or same length as following two tarsomeres combined; (1) longer than following two tarsomeres combined (length  =  29; CI  =  0.03; RI  =  0.67). See character 120 and table 3. Among the outgroup taxa, most species possess longer proximal tarsomeres (state 1) and only Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri possess shorter proximal tarsomeres (state 0).

122. Length of proximal metathoracic tarsomere: (0) shorter or same length as following two tarsomeres combined; (1) longer than following two tarsomeres combined (length  =  14; CI  =  0.07; RI  =  0.50). See character 120 and table 3. Among the outgroup taxa, most species possess longer proximal tarsomeres (state 1) and only Mydidae, with the exception of Mitrodetus dentitarsis, Neorhaphiomidas sp. (Mandurah), Opomydas townsendi, and Rhaphiomidas maehleri possess shorter proximal tarsomeres (state 0).

123. Pulvilli: (0) absent; (1) present (length  =  3; CI  =  0.33; RI  =  0.77). Martin (1968) observed that the pulvilli are absent in all Leptogastrinae and a few other species of Asilidae and concluded that the lack of pulvilli is a plesiomorphic character. Pulvilli are found here to be absent in all Leptogastrinae as well as Ablautus coquilletti, Acnephalum cylindricum, and Sisyrnodytes sp. (Gamka) (all Willistonininae) (state 0). Among the outgroup taxa, pulvilli are present in all species (state 1).

124. Number of dorsal longitudinal ridges on pulvilli: (0) one; (1) two (length  =  2; CI  =  0.50; RI  =  0.90). Bequaert (1961) and later Yeates and Irwin (1996) discussed the number of dorsal longitudinal ridges on the pulvilli in Asiloidea and found that all Apioceridae and Mydidae possess only a single ridge whereas Asilidae and other Asiloidea possess two. A single ridge is found here in Megapoda labiata and Pseudorus distendens (Dasypogoninae) as well as all Apioceridae and Mydidae (state 0). All remaining taxa possess two ridges (state 1). All species lacking pulvilli (see character 123) are coded as inapplicable (-).

125. Shape of claws: (0) abruptly angled distally; (1) fairly straight throughout (length  =  4; CI  =  0.25; RI  =  0.70). In Asiloidea, the claws are usually abruptly angled distally (state 0), but in all Leptogastrinae; Lestomyia fraudiger (Dasypogoninae), and Ablautus coquilletti, Acnephalum cylindricum, and Sisyrnodytes sp. (Gamka) (all Willistonininae), the claws are more or less straight and only slightly angled distally (state 1). Among the outgroup taxa, the claws are always angled distally (state 0).

126. Shape of tip of claws: (0) pointed; (1) blunt (length  =  3; CI  =  0.33; RI  =  0.33). The tip of the claws is usually pointed in Asiloidea (state 0), but Eichoichemus pyrrhomystax and Megaphorus pulchrus (Asilinae) as well as Lamyra gulo and Stiphrolamyra angularis (Laphriinae) possess claws with a blunt tip (state 1). Hull (1962), in delimiting the Ctenotini (Laphriinae), mentioned the presence of blunt claws. Among the outgroup taxa, the claws are always pointed (state 0).

127. Development of setiform empodium: (0) well developed (as long as claws); (1) reduced in length to varying degree; (2) minute or entirely absent (length  =  31; CI  =  0.06; RI  =  0.58). In all Heterodactyla the empodium is setiform (McAlpine, 1981; Stuckenberg, 2001), and in most Asilidae it is well developed and as long as the claws (state 0). In many Brachyrhopalinae; Dasypogoninae; Leptogastrinae; Trigonomiminae; and Stenopogoninae, the empodium is reduced in length to a varying degree (state 1). In Acronyches maya (Leptogastrinae) and Ablautus coquilletti (Willistonininae) it is reduced to a minute sclerite or entirely absent (state 2). Among the outgroup taxa, in all Apioceridae, Mydidae, and Scenopinidae the empodium is minute or entirely absent (state 2, see also Yeates and Irwin, 1996). Nemestrinidae is coded as inapplicable (-) as it does not possess a setiform empodium, which is an autapomorphy for the Heterodactyla (see Introduction).

Wing

Figures 73–85 illustrate a diversity of robber-fly wings and all wing veins and cells are labeled in figures 73–74.

Figure 73–74

Wings. 73. Hoplistomerus nobilis (Laphriinae). 74. Promachus amastrus (Asilinae). Wing veins and cells labeled; for abbreviations see character descriptions and McAlpine (1981). Scale lines  =  1 mm.

Figure 75–85

Wings of Asilidae species. 75. Laxenecera albicincta; 76. Laphystia tolandi; 77. Atomosia puella (all Laphriinae); 78. Philonicus albiceps; 79. Pogonioefferia pogonias; 80. Megaphorus pulchrus (all Asilinae); 81. Microstylum sp. (Karkloof) (Stenopogoninae); 82. Dasypogon diadema (Dasypogoninae); 83. Dioctria atricapillus (Dioctriinae); 84. Holcocephala calva (Trigonomiminae); 85. Euscelidia pulchra (Leptogastrinae). Scale lines  =  1 mm.

128. Anterior margin of wing in males: (0) straight; (1) C bulging anteriorly (length  =  5; CI  =  0.20; RI  =  0.00). In general, the anterior margin of the wing is straight in male and female Asilidae. In Clephydroneura sundaica and Nomomyia murina (Asilinae) as well as Emphysomera conopsoides, Ommatius incurvatus, and Ommatius tibialis (all Ommatiinae), the anterior margin of the wing is expanded anteriorly in males (state 1). Among the outgroup taxa, the anterior wing margin is always straight (state 0).

129. Upper calypter: (0) reduced; (1) well developed (length  =  2; CI  =  0.50; RI  =  0.75). The upper calypter is well developed in the Apsilocephalidae, Poecilognathus sp. (El Hacha) (Bombyliidae), Scenopinidae, and Therevidae (state 1), but reduced in all Apioceridae, Asilidae, and Mydidae (state 0). Yeates (1994) interpreted the reduction of the upper calypter differently and stated that it is reduced only in taxa with an elongated body shape, e.g., Leptogaster. It is certainly further reduced in Leptogastrinae than in other Asilidae, but in all examined species of Apioceridae, Asilidae, and Mydidae it is only a small lobe more or less continuous with the thorax along the axillary cord in contrast to other outgroup taxa in which it is a distinct lobe.

130. Alula: (0) well developed; (1) reduced in size to small lobe; (2) reduced entirely, nearly straight wing margin (length  =  16; CI  =  0.12; RI  =  0.72). Karl (1959) and Martin (1968) discussed the reduction of the alula as an autapomorphy of the Leptogastrinae in contrast to most other species of Asilidae, which have a well-developed alula. Two different states of reduction are recognized here in that the alula can be reduced, but a small lobe is still distinct (state 1; figs. 76–77, 83–84) and the alula is completely reduced so that the posterior wing margin is nearly straight (state 2; fig. 85). State 1 is found in most Brachyrhopalinae with the exception of Ceraturgus fasciatus, Chrysopogon pilosifacies, Codula limbipennis, and Dasycyrton gibbosus, as well as Cyrtophrys attenuatus and Deromyia fuscipennis (Dasypogoninae); Dicolonus simplex, Dioctria species, Echthodopa pubera, Eudioctria albius, and Nannodioctria sp. (formicaphaga) (all Dioctriinae); Atomosia puella, Cerotainia albipilosa, and Psilocurus modestus (all Laphriinae); most Trigonomiminae except Damalis species; Trichoura sp. (Tierberg) and Willistonina bilineata (Willistonininae); and Coleomyia setigera. State 2 is found in all Leptogastrinae and Stichopogoninae; Broticosia paramonovi (Dioctriinae); Goneccalypsis argenteoviridis and Loewinella nigripes (Laphriinae); and Oligopogon sp. (Cape Recife). Among the outgroup taxa, the alula is partly reduced in Apsilocephalidae (state 1).

131. Cell r₁: (0) open; (1) closed, R₁ and R₂₊₃ meet apically and form a stalk vein (petiolate); (2) closed, R₁ and R₂₊₃ meet directly at C (length  =  8; CI  =  0.25; RI  =  0.92). Hull (1962), Papavero (1973a), and subsequent authors widely used the development of cell r₁ for separating taxa of Asilidae, but highlighted the fact that this character has to be used with caution because a closed cell r₁ is supposed to have arisen several times independently (see Discussion). An open cell r₁, in which veins R₁ and R₂₊₃ reach the wing margin separately, is present in most Asilidae (state 0; figs. 81–85). In Asilinae; Megapoda labiata, Pseudorus distendens, and Thereutria amaraca (all Dasypogoninae); Laphriinae except Psilocurus modestus; Ommatiinae; as well as Apioceridae and Mydidae cell r₁ is closed (state 1 and 2). When cell r₁ is closed, a stalk vein formed by R₁ and R₂₊₃ is generally present (state 1; figs. 75, 77–80) and only in a few species of Laphriinae, i.e., Hoplistomerus nobilis, Laphystia species, Perasis transvaalensis, Protometer sp. (El Tuparro), and Zabrops tagax, R₁ and R₂₊₃ reach the wing margin at the same point (state 2; figs. 73, 76). Among the outgroup taxa, cell r₁ is closed in all Apioceridae and Mydidae (state 1) and open in the remaining species (state 0).

132. Cell r₄: (0) open; (1) closed (length  =  1; CI  =  1.00; RI  =  1.00). Cell r₄ is open in all Asilidae species (state 1). Among the outgroup taxa, it is closed only in Mydidae except Rhaphiomidas maehleri (state 1).

133. Cell r₄ shape: (0) R₄ and R₅ more or less parallel; (1) R₄ and R₅ diverging from each other; (2) R₄ and R₅ approximating each other medially; (3) R₄ and R₅ approximating each other distally (medially farthest apart); (4) R₄ and R₅ farthest apart in proximal one-third (length  =  26; CI  =  0.15; RI  =  0.69). R₄ and R₅ run more or less parallel in many Brachyrhopalinae; Cerotainia albipilosa (Laphriinae); Leptogastrinae; Ommatiinae; Holcocephala species and Trigonomima sp. (anamaliensis) (Trigonomiminae) (state 0; figs. 84–85). Within the Asilinae, the shape of cell r₄ varies considerably and states (0–3) are found with the most common configuration being a medially approximating R₄ and R₅ (state 2; fig. 78). State 3 is illustrated in figure 79. In all remaining species of Asilidae, R₄ and R₅ diverge from each other at the wing margin (state 1; figs. 73–77, 80–83). Among the outgroup taxa, Apioceridae and Apsilocephalidae possess state 0, Mydidae and Nemestrinidae possess state 4, and the remaining taxa exhibit state 1.

134. Cell r₅: (0) open; (1) closed (length  =  13; CI  =  0.07; RI  =  0.14). Cell r₅ is generally open among species of Asilidae (state 0, e.g.; fig. 74) and only in Apoclea algira (Asilinae); Hoplistomerus nobilis, Hyperechia nigripennis, Laxenecera albicincta, Nusa infumata, Perasis transvaalensis, Pilica formidolosa, Trichardis effrena, and Zabrops tagax (all Laphriinae); Daspletis stenoura (Stenopogoninae); Sisyrnodytes sp. (Gamka) (Willistonininae) cell r₅ is closed (state 1; figs. 73, 75). Among the outgroup taxa, cell r₅ is mostly open (state 0), but a few Mydidae possess a closed cell r₅ (state 1).

135. Cell m₃: (0) open; (1) closed (length  =  11; CI  =  0.09; RI  =  0.85). Cell m₃ is closed in all Asilinae; Laphriinae; Ommatiinae; and a few other Asilidae species (state 1; figs. 73–82). Among the outgroup taxa, this cell is absent due to the reduction of vein M₃ in Bombyliidae and the species are coded as inapplicable (-), but it is open in Scenopinidae, in which vein M₂ is absent (state 0). In all remaining species it is closed (state 1).

136. Cell cup: (0) open; (1) closed (length  =  18; CI  =  0.05; RI  =  0.66). Karl (1959) used the open or closed cell cup in his phylogeny of Asilidae. Cell cup is closed in all Asilinae; Bathypogoninae; Laphriinae; Ommatiinae; Phellinae; Stichopogoninae; many Stenopogoninae; and most Trigonomiminae (state 1; figs. 73–82). All remaining Asilidae possess an open cell cup (state 0; fig. 85). Trigonomima sp. (anamaliensis) is coded as polymorphic (state 0/1) because different individuals are observed, in which cell cup is open or closed. Among the outgroup taxa, cell cup is usually closed (state 1) except in Bombylius major (Bombyliidae) and Nemestrinidae.

137. Cell d, distal end: (0) base of M₂ and m-m closing cell in nearly a single line; (1) only m-m closing cell, M₁ branching proximally; (2) only base of M₂ closing cell, M₃ branching proximally; (3) M₂ and M₃ closing cell, not aligned; (4) only base of M₂ closing cell, m-m absent (or at least highly reduced); (5) only base of M₂ closing cell, distal M₃ absent (length  =  19; CI  =  0.26; RI  =  0.74). Cell d lies in the center of the wing in Asilidae. The branching of veins M₁ and M₃ forms the proximal angle. The distal tip is composed of veins M₁, M₂, and crossvein m-m (fig. 86). The branching between M₁ and M₂ varies in its position. State 0 is found when the branching is situated close to the distal tip of cell d and the proximal part of M₂ and crossvein m-m form nearly a straight line (figs. 83, 86). This arrangement is found in some species of Brachyrhopalinae; Dioctriinae; Leptogastrinae; Stichopogoninae; Trigonomiminae; and Willistonininae (state 0). State 1 is observed when the branching is situated more closely to the center of cell d, thus proximal part of M₂ and crossvein m-m are in an angle of about 90° (figs. 84, 86) and this arrangement is present in Lasiocnemus lugens, Leptogaster species, and Tipulogaster glabrata (Leptogastrinae). Laphriinae, with the exception of most Atomosiini; as well as Asilinae; Bathypogoninae; Dasypogoninae; Ommatiinae; Phellinae; Stenopogoninae except Connomyia varipennis; Tillobromatinae; and a few other species exhibit state 2 where M₃ branches off proximal to where M₂ originates and therefore m-m and M₂ are arranged in an angle (figs. 75–76, 78–82, 86). In the Atomosiini except Smeryngolaphria numitor (Laphriinae) m-m is absent or largely reduced so that the base of M₂ and part of M₃ are aligned in a straight line from anterior to posterior (state 4; figs. 77, 86; see character 150). In Townsendia albomacula (Stichopogoninae) and Trigonomima sp. (anamaliensis) (Trigonomiminae) the distal part of M₃ is absent and cell d is only closed by M₂ because m-m is also absent (state 5; fig. 86). Among the outgroup taxa, the Bombyliidae is coded as inapplicable (-) because vein M₃ is not present and therefore no cell d is developed. Mydidae and Nemestrinidae exhibit state 3 whereas Apioceridae and Prorates sp. (Escalante) exhibit state 2.

Figure 86

Schematic drawing of closure of cell d coded in character 137. Character states, wing veins, and wing cells labeled; character state 3, which is only found in Mydidae and Nemestrinidae, not illustrated; for abbreviations see character description and McAlpine (1981); not to scale.

138. Position of r-m with respect to cell d (or cell dm): (0) proximal half; (1) distal half; (2) distal tip (length  =  36; CI  =  0.05; RI  =  0.50). Yeates and Irwin (1996) used this character in the phylogenetic analysis of Apioceridae and Mydidae and established the position of r-m at the distal tip of cell d as an autapomorphy for a subordinate taxon of Mydidae. Within Asilidae and all remaining outgroup taxa the position of r-m with respect to cell d varies considerably and both states (0 and 1) are found throughout the taxa. State 0 is illustrated in figures 73–76 and 83–85 and state 1 is shown in figures 77 and 82. State 2 is here found to be present in Mydidae except Neorhaphiomidas sp. (Mandurah).

139. Development of discal medial cell (cell dm): (0) absent; (1) present (length  =  2; CI  =  0.50; RI  =  0.00). McAlpine (1981) and Yeates (1994) pointed out that with the loss of vein M₃, cell d cannot form. Vein M₃ is absent in Bombyliidae and therefore only cell dm is present in this taxon (state 1).

140. Extent of C: (0) around entire wing (circumambient); (1) terminating at wing apex; (2) terminating at apices of CuA₂ and A₁; (3) terminating at apex of M2; (4) terminating at apex of M1 (length  =  19; CI  =  0.21; RI  =  0.37). In most Asilidae and outgroup species, the costal vein C is developed around the entire wing (circumambient, state 0; figs. 74, 78–79, 81, 83–85). Different reductions of this configuration can be found in Laphriinae. In Laxenecera albicincta, Nusa infumata, and Trichardis effrena, C is reduced and terminates at the wing apex (state 1; fig. 75), whereas in Hoplistomerus nobilis, Laphystia species, and Protometer sp. (El Tuparro), it terminates at CuA₂ and A₁ (state 2; figs. 73, 76). In Aphestia annulipes, Atomosia puella, and Cerotainia albipilosa, C terminates at M₂ (state 2; fig. 77). Outside of Laphriinae, state 1 is found in Megaphorus pulchrus (fig. 80) and Satanas gigas (Asilinae) as well as Sisyrnodytes sp. (Gamka) (Willistonininae) and state 2 is found in Afroholopogon peregrinus, Dasycyrton gibbosus, and Rhabdogaster pedion (Brachyrhopalinae); Damalis species (Trigonomiminae); and Acnephalum cylindricum and Trichoura sp. (Tierberg) (Willistonininae). Among the outgroup taxa, Scenopinidae and Hemigephyra atra (Therevidae) possess a costal vein terminating at the wing apex (state 1) and in the Mydidae species Afroleptomydas sp. (Clanwilliam), Mitrodetus dentitarsis, Nemomydas brachyrhynchus, Neorhaphiomidas sp. (Mandurah), and Pseudonomoneura hirta, C terminates at vein M₁ (state 4).

141. Microtrichia on posterior wing margin: (0) arranged in a single plane; (1) arranged in two divergent planes; (2) absent (length  =  26; CI  =  0.07; RI  =  0.46). Londt (2005b) used the development of microtrichia on the posterior wing margin as a character to separate certain Afrotropical Apocleinae genera (sensu Papavero, 1973a) following Blasdale (unpublished) who has studied the Afrotropical Bactria Meigen, 1820, and Promachus species (see Geller-Grimm, 2004, for synonymy of these genera). In most taxa examined, the microtrichia are arranged in a single plane (state 0). In a number of Asilinae; Dasypogoninae; Laphriinae; as well as Acronyches maya (Leptogastrinae); Michotamia aurata (Ommatiinae); Phellus olgae (Phellinae); and Gonioscelis ventralis and Microstylum sp. (Karkloof) (Stenopogoninae), the microtrichia are arranged in two divergent planes (state 1). In Blepharotes splendidissimus and Satanas gigas (Asilinae); Obelophorus terebratus (Phellinae); and Sisyrnodytes sp. (Gamka) (Willistonininae), the microtrichia on the posterior margin of the wing are reduced (state 2). Among the outgroup taxa, Phycus frommeri (Therevidae) exhibits state 1 and in Mitrodetus dentitarsis and Neorhaphiomidas sp. (Mandurah) (Mydidae) the microtrichia are reduced (state 2).

142. Apex of R₂₊₃: (0) relatively straight; (1) arching sharply anteriorly in 90° angle (length  =  7; CI  =  0.14; RI  =  0.50). Wood (1981) and later authors used the sharply arching vein R₂₊₃ as a diagnostic feature to separate genera of Laphriinae from Laphystia and relatives (Laphystiinae sensu Papavero, 1973a). Most Asilidae exhibit a relatively straight vein R₂₊₃ (state 0; figs. 74–75, 77–81, 83–85). Among the Laphriinae, Hoplistomerus nobilis, Perasis transvaalensis, Protometer sp. (El Tuparro), Trichardis effrena, and Zabrops tagax; as well as Bathypogoninae; Gonioscelis ventralis, Ospriocerus aeacus, Scleropogon subulatus, and Stenopogon species (all Stenopogoninae); Hypenetes stigmatias (Tillobromatinae); and Sisyrnodytes sp. (Gamka) (Willistonininae) possess a sharply anteriorly arching vein R₂₊₃ (state 1; fig. 73). Among the outgroup taxa, the R₂₊₃ is relatively straight (state 0).

143. Apex of R₄: (0) terminating anterior to wing apex; (1) terminating at wing apex; (2) terminating posterior to wing apex; (3) not reaching C (or wing margin) (length  =  14; CI  =  0.21; RI  =  0.60). In most Asilidae species, the apex of vein R₄ reaches the wing margin anterior to the wing apex (state 0; figs. 73–84). In Austrosaropogon nigrinus, Leptarthrus brevirostris, and Rhabdogaster pedion (all Brachyrhopalinae); Cyrtophrys attenuatus, Deromyia fuscipennis, Molobratia teutonus, and Senobasis corsair (all Dasypogoninae); Dioctria hyalipennis, Dioctria rufipes, and Nannodioctria sp. (formicaphaga) (all Dioctriinae); Euscelidia pulchra (Leptogastrinae); Lasiopogon species, Lissoteles aquilonius, Stichopogon punctum, and Stichopogon trifasciatus (all Stichopogoninae); and Trigonomima sp. (anamaliensis) (Trigonomiminae), R₄ reaches the wing margin at the wing apex (state 1). In Leptogastrinae except Euscelidia pulchra and Pegesimallus laticornis (Dasypogoninae), vein R₄ reaches the wing margin always posterior to the wing apex (state 2; fig. 85). Among the outgroup taxa, vein R₄ does not reach the costa or the wing margin in Mydidae except Rhaphiomidas maehleri (state 3).

144. Shape of R₄: (0) relatively straight; (1) strongly sinuate (length  =  7; CI  =  0.14; RI  =  0.40). Papavero (1973a) established the strongly sinuate vein R₄ (state 1) as the defining characteristic to separate Laphriinae and a number of genera related to Laphystia (Laphystiinae sensu Papavero, 1973a). Most species of Asilidae exhibit a relatively straight vein R₄ (state 0; figs. 74–75, 77–85). State 1 is found in the following species of Laphriinae, i.e., Hoplistomerus nobilis, Laphystia species, Perasis transvaalensis, Protometer sp. (El Tuparro), Stiphrolamyra angularis, Trichardis effrena, and Zabrops tagax, as well as Bathypogoninae and Stenopogon rufibarbis (Stenopogoninae) (figs. 73, 76). Among the outgroup taxa, R₄ is sinuate only in Bombylius major (Bombyliidae) (state 1).

145. R₅ apex: (0) terminating anterior to wing apex; (1) terminating posterior to wing apex; (2) not reaching C (or wing margin) (length  =  6; CI  =  0.33; RI  =  0.69). Woodley (1989) postulated the strongly anteriorly curved veins R₅ and M₁ that reach the wing margin anterior to the wing apex as a synapomorphy for a taxon composed of Apioceridae plus Mydidae. Within Asilidae R₅ usually reaches the wing margin posterior to the wing apex (state 1; figs. 73–78, 80–85), but in Lochmorhynchus albicans, Pogonioefferia pogonias, Proctacanthus philadelphicus, and Satanas gigas (all Asilinae) R₅ is curving anteriorly so that vein R₅ terminates anterior to the wing apex (state 0; fig. 79). Among the outgroup taxa, R₅ terminates anterior to the wing apex in Apioceridae, Rhaphiomidas maehleri (Mydidae), and Nemestrinidae (state 0), but in all other Mydidae species, vein R₅ does not reach the costa vein or the wing margin (state 2). In Apsilocephalidae, Bombyliidae, Scenopinidae, and Therevidae R₅ terminates posterior to the wing apex (state 1).

146. Stump vein (supernumerary crossvein) on R₄: (0) absent; (1) present, short stump vein not reaching R₂₊₃; (2) present, long stump vein reaching base of R₂₊₃ (length  =  11; CI  =  0.18; RI  =  0.35). The wing venation in Asilidae is relatively straightforward and simple compared to, for example, Mydidae or Nemestrinidae. The short stump vein present at the base of R₄ just after departing from the common R₄₊₅ in a few taxa is, however, an interesting case of the development of an additional vein not commonly present in other taxa. Oldroyd (1974b) and Londt (2005b) discussed the different configurations of this supernumerary vein in Afrotropical taxa. Hennig (1954) postulated that this vein represents R₃ and illustrated its presence in Promachus leoninus Loew, 1848, and related genera (Asilinae) as well as Apiophora paulseni Philippi, 1865 (as Rhopalia, Mydidae). This is only consequent as he postulated that R₃ is usually lost in Asiloidea, so that the second radial vein is R₂ and not R₂₊₃ as interpreted by, for example, McAlpine (1981: fig. 4.32). Papavero (1973a) followed Hennig's view and labeled the stump vein as R₃. Most Asilidae species lack any sign of the stump vein on R₄ (state 0; figs. 73, 75–78, 81–85). It is found consistently in Asilinae, e.g., Apoclea algira, Nomomyia murina, and Pogonioefferia pogonias, and is also found in Obelophorus terebratus and Phellus olgae (Phellinae) and Acnephalum cylindricum (Willistonininae) (state 1; fig. 79). A special case is exhibited in Eichoichemus pyrrhomystax, Megaphorus pulchrus, Philodicus tenuipes, and Promachus amastrus (all Asilinae), in which the stump vein connects R₂₊₃ and R₄ and forms a supernumerary crossvein thereby dividing cell r₃ into two cells r_3a and r_3b (state 2; figs. 74, 80). Among the outgroup taxa, a short stump vein is found in Apsilocephalidae, Poecilognathus sp. (El Hacha) (Bombyliidae), Afroleptomydas sp. (Clanwilliam), Nemomydas brachyrhynchus, Opomydas townsendi, and Pseudonomoneura hirta (all Mydidae) (state 1).

147. Apex of M₁: (0) terminating anterior to wing apex; (1) terminating posterior to wing apex; (2) not reaching C (or wing margin) (length  =  4; CI  =  0.50; RI  =  0.80). Woodley (1989) postulated the anteriorly curved veins R₅ and M₁ as a synapomorphy for a taxon composed of Apioceridae plus Mydidae (see character 145). In Asilidae, vein M₁ terminates posterior to the wing apex (state 1; figs. 73–85). Among the outgroup taxa, M₁ always curves anteriorly to terminate anterior to the wing apex in Apioceridae, Mydidae except Pseudonomoneura hirta, and Nemestrinidae (state 1). In Satanas gigas (Asilinae) and Pseudonomoneura hirta (Mydidae), vein M₁ does not reach the costa vein or the wing margin (state 2).

148. Fusion of M₁ and M₂: (0) separated; (1) fused (length  =  2; CI  =  0.50; RI  =  0.85). Yeates and Irwin (1996) used the fusion of M₁ or M₁+M₂ with R₄ as a character in the phylogenetic analysis of Apioceridae and Mydidae. Here, the fusion of M₁ and M₂ itself is considered and this state is found in Mydidae except Rhaphiomidas maehleri and Nemestrinidae (state 1). All Asilidae species possess separated M₁ and M₂ veins (state 0; figs. 73–85).

149. Development of M₃ (distal section beyond m-m): (0) absent; (1) present (length  =  4; CI  =  0.25; RI  =  0.25). Yeates (1994) mentioned the reduction of vein M₃ in Bombyliidae and Scenopinidae and this is confirmed here (state 0). In Townsendia albomacula (Stichopogoninae) and Trigonomima sp. (anamaliensis) (Trigonomiminae), vein M₃ is also reduced beyond crossvein m-m (state 0; fig. 86). All remaining Asilidae species exhibit a well-developed vein M₃ (state 1; figs. 73–85).

150. Alignment of base of M₂ and m-m: (0) not aligned; (1) aligned from anterior to posterior; (2) aligned from proximal to distal (nearly parallel to posterior margin of wing) (length  =  4; CI  =  0.50; RI  =  0.71). Macquart (1838), in describing the genus Atomosia and later Hermann (1912) in diagnosing a taxon comprising Atomosia and related taxa from South America, highlighted the distinct wing venation of vein M₂ and crossvein m-m in these species. In most species of Asilidae, M₂ and m-m are not aligned in any particular way (state 0; figs. 73–76, 78–80, 82–85). In Aphestia annulipes, Atomosia puella, Cerotainia albipilosa, Goneccalypsis argenteoviridis, and Loewinella nigripes (all Laphriinae), M₂ and m-m are aligned from anterior to posterior (state 1; fig. 77). Orthogonis scapularis (Laphriinae) is coded here to be polymorphic (states 0/1) because some specimens exhibit a clear alignment of the veins (state 1) and in others, hardly any alignment is found (state 0). In Archilestris magnificus (Dasypogoninae); Obelophorus terebratus and Phellus olgae (Phellinae); and Microstylum sp. (Karkloof) (Stenopogoninae), M₂ and m-m are also aligned, but in a line from proximal to distal and running nearly parallel to the posterior wing margin (state 2; fig. 81). Among the outgroup taxa, these two veins are always unaligned (state 0).

Abdomen (Female and Male)

151. Sculpture of tergites: (0) smooth, setae with small socket only; (1) indented, setae in large grooves (length  =  5; CI  =  0.20; RI  =  0.60). In general, the surface of the tergites is smooth and setal sockets are only barely visible under a stereomicroscope (state 0). As has been pointed out by Hermann (1912) for some South American Laphriinae, the setal sockets can be very large and distinct. This is found in Chrysopogon pilosifacies, Codula limbipennis, and Leptarthrus brevirostris (all Brachyrhopalinae), as well as Aphestia annulipes, Atomosia puella, Cerotainia albipilosa, Goneccalypsis argenteoviridis, Hoplistomerus nobilis, Loewinella nigripes, Stiphrolamyra angularis, and Trichardis effrena (all Laphriinae) (state 1). Among the outgroup taxa, the tergites are always smooth (state 0).

152. T1 medially: (0) entirely sclerotized medially; (1) divided medially by membraneous area into two halves (length  =  12; CI  =  0.08; RI  =  0.38). T1 is usually composed of a single, narrow sclerite with an entire posterior margin (state 0). A few Asilidae exhibit a medially divided tergite (state 1), which is indicated by an indentation of the posterior margin. The following species exhibit a divided T1: Blepharotes splendidissimus, Eichoichemus pyrrhomystax, Heligmonevra laevis, Lochmorhynchus albicans, Megaphorus pulchrus, Myaptex brachyptera, Neoitamus cyanurus, Philodicus tenuipes, Proctacanthus philadelphicus, Promachus amastrus (all Asilinae); as well as Megapoda labiata (Dasypogoninae); Holcocephala abdominalis, Rhipidocephala sp. (Harold Johnson), and Trigonomima sp. (anamaliensis) (all Trigonomiminae); and Willistonina bilineata (Willistonininae). Among the outgroup taxa, Bombyliidae and Scenopinidae exhibit a divided T1 (state 1).

153. T2 length: (0) wider than long or as wide as long; (1) longer than wide, less than twice as long; (2) longer than wide, more than twice as long (length  =  18; CI  =  0.11; RI  =  0.44). Martin (1968) established the elongated second abdominal tergite in Leptogastrinae species as an autapomorphy for this taxon and it has since been used in identification keys by several authors (e.g., Wood, 1981). In most Asilidae, T2 is wider than long (state 0). In Clephydroneura sundaica and Lycomya germainii (Asilinae); Nannocyrtopogon lestomyiformis, Nicocles politus, and Rhabdogaster pedion (all Brachyrhopalinae); Cyrtophrys attenuatus, Deromyia fuscipennis, and Pegesimallus laticornis (all Dasypogoninae); Broticosia paramonovi, Dioctria hyalipennis, Dioctria rufipes, and Nannodioctria sp. formicaphaga (all Dioctriinae); Lamyra gulo (Laphriinae); Ommatiinae except Cophinopoda chinensis; Plesiomma sp. (Guanacaste) and Stenopogon rufibarbis (Stenopogoninae); Tillobroma punctipennis (Tillobromatinae); Holcocephala calva (Trigonomiminae); and Oligopogon sp. (Cape Recife), T2 is wider than long, but less than twice as long (state 1). All Leptogastrinae species exhibit state 2. Among the outgroup taxa, T2 is always wider than long (state 0).

154. T2 shape dorsally: (0) dorsal surface smooth, without protuberance; (1) dorsal surface with proximal protuberance (length  =  4; CI  =  0.25; RI  =  0.76). In general the dorsal surface of T2 is flat or only slightly elevated medially and declining toward the posterior margin in Asilidae (state 0). In a number of species, a proximal protuberance is evident that is distinct and easily observed (state 1). Papavero (1975) called these protuberances on T2 bullae, but these simple structures are not homologous to the bullae found in some Mydidae species, which are situated at the posterior margin of T2 and separated medially (see character 155). The following species exhibit state 1: Ceraturgus fasciatus (Brachyrhopalinae); Cyrtophrys attenuatus, Deromyia fuscipennis, Megapoda labiata, Molobratia teutonus, Pegesimallus laticornis, and Pseudorus distendens (all Dasypogoninae); Dioctria species, Echthodopa pubera, Eudioctria albius, Myelaphus melas, and Nannodioctria sp. (formicaphaga) (all Dioctriinae). Among the outgroup taxa, all species lack dorsal protuberances on T2 (state 0).

155. Bullae on T2: (0) absent; (1) present (length  =  1; CI  =  1.00; RI  =  1.00). Yeates and Irwin (1996) postulated the bullae on the lateroposterior margin of T2 as an autapomorphy of a subordinate taxon of Mydidae. It is found here in all representatives of Mydidae, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri (state 1).

156. T2–3 development: (0) tergites square to rectangular (parallel-sided); (1) tergites forming a distinct “waist” (length  =  6; CI  =  0.16; RI  =  0.50). In general, the abdomen is more or less parallel-sided, sometimes becoming wider toward the apex to accommodate the genitalia in Asilidae (state 0). Hennig (1973) mentioned a proximally constricted abdomen in Asilidae as an autapomorphy and attributed this character to Hull (1962; note: I am unable to find any reference to such a character in Hull's study). In some Asilidae species, namely Brachyrhopala ruficornis (Brachyrhopalinae); Blepharepium cajennensis, Cyrtophrys attenuatus, Molobratia teutonus, and Pegesimallus laticornis (all Dasypogoninae); Nannodioctria sp. (formicaphaga) (Dioctriinae); Lamyra gulo (Laphriinae); and Emphysomera species and Michotamia aurata (Ommatiinae), T2–3 are narrower than the following tergites and therefore the abdomen appears constricted proximally and waistlike (state 1). Among the outgroup taxa, the proximal abdomen is never waistlike (state 0). In some instances, it is difficult to decide whether the constriction is an artifact of a dried specimen and this is especially true in females, which have an expanded distal abdomen due to the large number of eggs held within it.

157. T3–6 development: (0) tergites entirely sclerotized; (1) tergites unsclerotized medially (length  =  2; CI  =  0.50; RI  =  0.00). Abdominal tergites are usually sclerotized entirely (state 0), but two species of Trigonomiminae, i.e., Rhipidocephala sp. (Harold Johnson) and Trigonomima sp. (anamaliensis), exhibit medially unsclerotized tergites (state 1). Among the outgroup taxa, T3–6 are always well developed and sclerotized (state 0).

158. T6 shape: (0) T6 the same as T5; (1) T6 cup-shaped, last visible tergite concealing remaining tergites (length  =  1; CI  =  1.00; RI  =  1.00). The abdominal tergites are very uniform in Asilidae, in that with the exception of the tergites immediately anterior to the genitalia, they are not modified and appear very similar throughout the abdomen. T5 and T6 are no exception and are usually of the same shape although T6 might be shorter than T5 (state 0). In Atomosia puella and Cerotainia albipilosa (Laphriinae), T6 is very large, cup-shaped, and the last apparent segment concealing the following tergites and genitalia (state 1). Among the outgroup taxa, T6 has always the same shape as T5 (state 0).

159. S2 development: (0) undivided medially; (1) divided medially into two equal halves separated by fenestra (length  =  1; CI  =  1.00; RI  =  1.00). Martin (1968) postulated the divided S2 with fenestra as an autapomorphy for his Leptogastridae. State 1 is found here in all Leptogastrinae species, but not in any other Asilidae. The fenestrae, small unsclerotized parts of the cuticle, are situated medially on the elongated second abdominal segment. It is important to highlight the observation that S2 also appears to be divided proximally in, for example, Ommatius or Damalis, but this state is not considered homologous to the state found in Leptogastrinae. The problem lies in the difficulty to distinguish between S1 and S2. The unsclerotized area in Damalis could be interpreted as the intersegmental membrane between S1 and S2 and not as a subdivision of S2. Among the outgroup taxa, S2 is never subdivided (state 0).

Female Abdomen

Note: Broticosia paramonovi (Dioctriinae), Emphysomera conopsoides (Ommatiinae), Ancylorhynchus fulvicollis (Stenopogoninae), and Nemomydas brachyrhynchus (Mydidae) are represented by male specimens only and could not be coded for the female abdominal characters. Only previously dissected specimens of Megapoda labiata (Dasypogoninae) where available for study and characters 166–174 and 181–185 could not be coded for this species. Neorhaphiomidas sp. (Mandurah) (Mydidae) and Prorates sp. (Escalante) (Scenopinidae) are not coded for characters 175–182. In some instances reference is made to illustrations of species represented in this study published previously by Theodor (1976) who studied the internal female genitalia in detail.

160. Lateral macrosetae on T2: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  7; CI  =  0.28; RI  =  0.88). The presence of lateral marginal macrosetae, also often called discal setae, on the abdominal tergites has been widely used in the identification of Asilinae species in which these macrosetae are often well developed. Many Laphriinae species also have lateral macrosetae on the abdominal tergites, but these setae are situated medially and not on the posterior margin of the tergites, therefore, being interpreted as nonhomologous to the marginal setae in Asilinae. Table 4 summarizes the presence/absence of marginal setae (state 1) and medial setae (state 2) in Asilinae and Laphriinae. Daspletis stenoura (Stenopogoninae) exhibits also marginal setae on T2. Among the outgroup taxa, neither marginal nor medial setae are developed (state 0).

Table 4

Presence/absence of marginal and medial macrosetae on abdominal tergites 2–7 in males and females of Asilinae and Laphriinae (chars. 160–165 and 186–191). Symbols: –  =  no macrosetae present; •  =  marginal macrosetae present; ▵  =  medial macrosetae present

161. Lateral macrosetae on T3: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  8; CI  =  0.25; RI  =  0.84). See character 160 and table 4.

162. Lateral macrosetae on T4: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  9; CI  =  0.22; RI  =  0.82). See character 160 and table 4. State 1 is also found in Emphysomera pallidapex, Afroestricus chiastoneurus, and Ommatius tibialis (all Ommatiinae).

163. Lateral macrosetae on T5: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  15; CI  =  0.13; RI  =  0.60). See character 160 and table 4. State 1 is also found in Emphysomera pallidapex, Afroestricus chiastoneurus, and Ommatius tibialis (all Ommatiinae).

164. Lateral macrosetae on T6: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  12; CI  =  0.16; RI  =  0.44). See character 160 and table 4. State 1 is also found in Emphysomera pallidapex, Afroestricus chiastoneurus, and Ommatius tibialis (all Ommatiinae).

165. Lateral macrosetae on T7: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  4; CI  =  0.50; RI  =  0.50). See character 160 and table 4. State 1 is also found in Emphysomera pallidapex, Afroestricus chiastoneurus, and Ommatius tibialis (all Ommatiinae).

166. Segment 7: (0) S7 and T7 normally developed; (1) S7 and T7 distinctly shortened; (2) S7 and T7 elongated; (3) only S7 reduced in size (length  =  3; CI  =  1.00; RI  =  1.00). The abdominal segments are usually of the same size and in some instances become slightly shorter only toward the apex of the abdomen. Only the segments that are directly involved in oviposition (see character 167) are altered in their proportions. In a few species, however, segment 7 is morphologically different from the following segments. In Aphestia annulipes, Atomosia puella, and Cerotainia albipilosa (all Laphriinae) segment 7 is much shorter than segment 8 (state 1) and in Leptarthrus brevirostris (Brachyrhopalinae) segment 7 is elongated (state 2). In Dicolonus simplex (Dioctriinae) S7 is reduced in size whereas T7 is normally developed (state 3). Among the outgroup taxa, segment 7 is always normally developed (state 0).

167. Segments comprising ovipositor: (0) 8th and following; (1) 7th and following (length  =  3; CI  =  0.33; RI  =  0.60). The specialized abdominal segments that are associated with the oviposition of eggs are generally referred to as the ovipositor. The ovipositor of Diptera is not homologous to the “true” ovipositor found in other Endopterygota (e.g., Mickoleit, 1973; McAlpine, 1981). In general, the ovipositor comprises segments 8 and 9 (state 0), but in Leptarthrus brevirostris (Brachyrhopalinae); Aphestia annulipes, Atomosia puella, and Cerotainia albipilosa (all Laphriinae); and Obelophorus terebratus and Phellus olgae (Phellinae), segment 7 is specialized and involved in oviposition (state 1). Among the outgroup taxa, segments 8 and following comprise the ovipositor (state 0).

168. Direction of setae on T8: (0) posteriorly; (1) anteriorly; (2) erect (length  =  29; CI  =  0.06; RI  =  0.65). T8 is generally modified and part of the ovipositor (see character 167). The setation of T8 varies among Asilidae species with respect to the orientation of the setae. All species of Dioctriinae; Laphriinae; Leptogastrinae; Ommatiinae; Trigonomiminae; and a few other species exhibit state 0 in that the setae are directed posteriorly in the same direction as setae on proximal tergites. In Dasycyrton gibbosus, Holopogon priscus, and Leptarthrus brevirostris (all Brachyrhopalinae); Dasypogon diadema, Diogmites grossus, Pegesimallus laticornis, Saropogon dispar, and Thereutria amaraca (all Dasypogoninae); Gonioscelis ventralis (Stenopogoninae); Stichopogon punctum (Stichopogoninae); Lycostommyia albifacies (Tillobromatinae); and Acnephalum cylindricum (Willistonininae), the setae are directed anteriorly (state 1). Many Asilinae; Bathypogoninae; Brachyrhopalinae; Phellinae; Stenopogoninae; Tillobromatinae; Willistonininae; and Oligopogon sp. (Cape Recife) possess erect setae, which are directed perpendicular to the surface (state 2). In some instances, states (1) and (2) are difficult to distinguish. Among the outgroup taxa, state 0 is found in Apsilocephalidae, Bombyliidae, Nemestrinidae, Scenopinidae, and Therevidae and only within Apioceridae and Mydidae are states (1) and (2) present.

169. T8 anterior rectangular projection: (0) absent, no apodeme present; (1) present, rectangular apodeme, entirely fused to T8; (2) present, rectangular apodeme, dissociated from T8; (3) present, cylindrical apodeme, fused to T8 (length  =  20; CI  =  0.15; RI  =  0.61). The anterior margin of T8 is generally simple with no particular appendages or projections (state 0). When macerated in KOH, a narrow projection, which is situated ventral of the intersegmental membrane, and fused entirely to T8 can be found in all Bathypogoninae; Tillobromatinae; Willistonininae; and a number of Brachyrhopalinae; Dasypogoninae; and Stenopogoninae (state 1; fig. 87). Senobasis corsair (Dasypogoninae) exhibits a projection as well, but it is not fused to T8 (state 2). Among the outgroup taxa, Bombylius major (Bombyliidae) and Hemigephyra atra (Therevidae) exhibit a smaller, cylindrical apodeme that is fused to T8 (state 3). Most Apioceridae and Mydidae species, with the exception of Mydas clavatus and Rhaphiomidas maehleri, which lack a projection, exhibit state 1.

Figure 87–89

Female posterior part of abdomen in dorsal view with tergites labeled. 87. Hypenetes stigmatias (Tillobromatinae). 88. Dasyllis funebris (Laphriinae), T10 grey. 89. Acronyches maya (Leptogastrinae), lightly sclerotized T10 grey. Apod  =  anterior apodeme of T8; vestiture omitted. Scale lines  =  1 mm.

170. Shape of S8: (0) platelike, slightly emarginate mediodistally; (1) platelike, with median area elongated; (2) proximally platelike, hypogynial valves forming keel distally; (3) keellike throughout; (4) platelike, hypogynial valves extending; (5) platelike, distal margin medially membranous; (6) platelike, hypogynial valves separated (surrounded by membrane); (7) platelike, divided into two halves; (8) S8 modified to form part of sand chamber (length  =  33; CI  =  0.24; RI  =  0.77). The shape of the ovipositor varies considerably in Asilidae as different oviposition sites are used by different species. Leptogaster cylindrica (Leptogastrinae) and Holcocephala calva (Trigonomiminae) perch on vegetation and drop their eggs to the ground whereas Pogonioefferia pogonias places the eggs between the leaves of grass stems and Laphria flava deposits eggs in old tree trunks. S8 is modified to accommodate these different oviposition mechanisms and figure 90 illustrates the different configurations of S8 that are here coded as character states. A simple platelike S8 with slight emarginations (state 0) is found in Leptogastrinae, with the exception of Acronyches maya (state 6); Ommatiinae; Trigonomiminae, with the exception of Rhipidocephala sp. (Harold Johnson) (state 5); some Laphriinae and most of the outgroup taxa (state 0; figs. 96, 98). State 1 is exhibited in Leptarthrus brevirostris (Brachyrhopalinae) and the following Laphriinae: Andrenosomatini, Ctenotini, Laphriini, Nusa infumata, Orthogonis scapularis, and Smeryngolaphria numitor (fig. 94). State 2 is found in Chrysopogon pilosifacies and Codula limbipennis (Brachyrhopalinae) and the following Asilinae species: Afromochtherus mendax, Asilus species, Apoclea algira, Blepharotes splendidissimus, Colepia rufiventris, Congomochtherus inachus, Heligmonevra laevis, Lycomya germainii, Megaphorus pulchrus, Neoitamus cyanurus, Neomochtherus pallipes, Nomomyia murina, Philodicus tenuipes, Philonicus albiceps, Proctacanthus philadelphicus, Promachus amastrus, Satanas gigas, and Zosteria rosevillensis (figs. 92, 95). All remaining Asilinae together with Pseudorus distendens (Dasypogoninae) and Obelophorus terebratus and Phellus olgae (Phellinae) possess an ovipositor that is keellike throughout (state 3; fig. 93). A platelike S8 with extending hypogynial valves (state 4) is found in all Bathypogoninae; Tillobromatinae; and in the majority of species of Brachyrhopalinae; Dasypogoninae; Dioctriinae; Stenopogoninae; and Willistonininae. All Stichopogoninae; Threnia carbonaria (Asilinae); Cophura brevicornis and Nicocles politus (Brachyrhopalinae); Dasypogon diadema, Diogmites grossus, and Thereutria amaraca (all Dasypogoninae) possess a platelike S8 in which the hypogynial valves are separated and surrounded by membrane (state 6; fig. 97). Connomyia varipennis (Stenopogoninae); Apiocera painteri (Apioceridae), and Therevidae have a platelike S8 that is divided into two halves (state 7). Bombylius major (Bombyliidae) has a modified ovipositor to form a sand chamber (see Yeates, 1994) and is coded here as state 8.

Figure 90

Schematic illustration of shape of female S8 in ventral view as coded in character 170. Numbers refer to character states; state 8, which is only found in Bombylius major (Bombyliidae), is illustrated in Yeates (1994); vestiture omitted; not to scale.

Figure 91–92

SEM micrographs of female ovipositor in lateral view with morphological features labeled. 91. Lasiopogon cinctus (Stichopogoninae). 92. Promachus amastrus (Asilinae). For abbreviations see Materials and Methods. Scale lines  =  200 µm.

Figure 93–98

SEM micrographs of female ovipositor in lateral and ventral view. 93. Dysmachus trigonus (Asilinae), lateral. 94. Laphria flava (Laphriinae), ventral. 95. Philonicus albiceps (Asilinae), lateral. 96. Dioctria atricapillus (Dioctriinae), ventral. 97. Lasiopogon cinctus (Stichopogoninae), ventral. 98. Euscelidia pallasii (Wiedemann, 1818) (Leptogastrinae), ventral. Scale lines  =  200 µm.

171. T9 shape (when distinguishable): (0) small, triangular sclerite; (1) wide, rectangular sclerite without median protuberance; (2) wide, rectangular sclerite with median protuberance (length  =  2; CI  =  1.00; RI  =  1.00). Yeates (1994) pointed out that in Apioceridae and Mydidae T9 bears a dorsal spine whereas other Asiloidea have a smooth dorsal surface. T9 and T10 are fused in many Asiloidea (see character 173) making it difficult to examine the shape of T9. All taxa with a fused T9 and T10 are coded as inapplicable (-). For all Asilidae in which both sclerites are distinguishable, T9 is small and triangular (state 0; figs. 87, 91). Among the outgroup taxa, state 1 is found in Bombylius major (Bombyliidae), Nemestrinidae, and Phycus frommeri (Therevidae) and state 2 is only found in Apioceridae and Mydidae.

172. T10 shape: (0) single sclerite; (1) tergite partly divided posteromedially; (2) tergite divided into two heavily sclerotized acanthophorite plates; (3) tergite divided into two heavily sclerotized sclerites that are far apart; (4) tergite divided into two slightly sclerotized sclerites; (5) T10 lost or fused to cerci (length  =  21; CI  =  0.23; RI  =  0.82). Hull (1962), Oldroyd (1964), and Hennig (1976) observed the division of T9 into two sclerites called acanthophorite plates that bear generally strong spurs used to excavate soil for oviposition. Hardy (1944) was ambiguous in his description and mentioned both T9 and T10 as bearing the spurs although he also indicated that the sclerites appear fused in many instances. It has been shown by Adisoemarto and Wood (1975), Irwin (1976), McAlpine (1981) and subsequent authors that it is actually T10 that is divided. T9 is often reduced to a small sclerite and sometimes even fused to T10 so that both are indistinguishable. The reduction of T9 can probably be attributed to the transformation of S9 into the furca (see character 184). A single, separate sclerite T10 is found only in the majority of species of Asilinae; Dioctriinae; and Ommatiinae (state 0; figs. 93, 95); all remaining Asilidae possess a modified T10. A partly posteromedially divided T10 is found in Megaphorus pulchrus, Nomomyia murina, Promachus amastrus, and Zosteria rosevillensis (all Asilinae) (state 1; fig. 92). A divided T10 with two heavily sclerotized acanthophorite plates is found in all Bathypogoninae; Brachyrhopalinae; Dasypogoninae except Senobasis corsair; Phellinae; Stenopogoninae; Stichopogoninae; Tillobromatinae; Willistonininae; and Oligopogon sp. (Cape Recife) (state 2; figs. 87, 91, 97). Species of Andrenosomatini (Laphriinae) exhibit state 3 in which two separate heavily sclerotized sclerites are positioned far apart from each other (fig. 88). In Leptogastrinae; Emphysomera pallidapex (Ommatiinae); as well as Holcocephala species and Trigonomima sp. (anamaliensis) (Trigonomiminae), T10 is visible only as two weakly sclerotized sclerites that are positioned far apart (state 4; fig. 89). Karl (1959) referred to a personal communication by E. Hennig who found that the 9th abdominal segment is absent in Leptogastrinae. This conclusion could be attributed to the difficulty of distinguishing the weakly sclerotized plates forming T10 in Leptogastrinae. T10 is either entirely reduced or fused to the cerci so that it is indistinguishable in most species of Laphriinae as well as Damalis species (Trigonomiminae) (state 5). Among the outgroup taxa, either a simple T10 is developed in Apsilocephalidae, Poecilognathus sp. (El Hacha) (Bombyliidae), and Hemigephyra atra (Therevidae) (state 0), or acanthophorite plates are developed as in Apioceridae, Bombylius major (Bombyliidae), Mydidae except Mydas clavatus, Scenopinidae, and Phycus frommeri (Therevidae) (state 2).

173. Fusion of T9 and T10: (0) not fused; (1) partly fused; (2) entirely fused, sclerites not distinguishable (length  =  12; CI  =  0.16; RI  =  0.83). An entirely separated T9 and T10 is found only in Chrysopogon pilosifacies, Codula limbipennis, Cophura brevicornis, and Nicocles politus (all Brachyrhopalinae) as well as Blepharepium cajennensis and Neodiogmites melanogaster (Dasypogoninae) (state 0). A partly fused T9 and T10 is found in most Brachyrhopalinae and Dasypogoninae; Phellinae; Stenopogoninae; Tillobromatinae; Willistonininae; and Oligopogon sp. (Cape Recife) (state 1; fig. 87). The two tergites are entirely fused or at least not distinguishable, which could also be due to the reduction of T9, in all Asilinae; Bathypogoninae; Leptarthrus brevirostris (Brachyrhopalinae); Senobasis corsair and Molobratia teutonus (Dasypogoninae); Dioctriinae; Laphriinae; Leptogastrinae; Ommatiinae; and Trigonomiminae (state 2; fig. 92). Among the outgroup taxa, Nemestrinidae exhibit state 0, Bombylius major (Bombyliidae) state 1, and Apioceridae, Apsilocephalidae, Poecilognathus sp. (El Hacha) (Bombyliidae), Mydidae, Scenopinidae, and Therevidae state 2.

174. Spurs on ovipositor: (0) absent; (1) acanthophorite spurs on T10; (2) spurs on cerci (length  =  16; CI  =  0.12; RI  =  0.81). Hardy (1927a, 1944), Hull (1962), and Papavero (1973a) discussed the presence of spurs on the modified T10 within Asilidae and Asiloidea. Many species of Asilidae lack spurs on T10 no matter how this tergite is modified (see also character 172). Even when T10 is divided into two acanthophorite plates, the large, movable spurs are not always present (note: I use the term spurs instead of spines as spurs have been defined by Torre-Bueno [Nichols, 1989] as moveable structures connected to the cuticula by a joint whereas spines are not connected by a joint). Spurs are absent in most Asilinae; Dioctriinae; Laphriinae; Leptogastrinae; Ommatiinae; Trigonomiminae; and a few other species (state 0; figs. 88–89, 92, 93). Acanthophorite spurs are present in all Bathypogoninae; most Brachyrhopalinae and Dasypogoninae; Phellinae; Stenopogoninae; Tillobromatinae; Willistonininae; and Oligopogon sp. (Cape Recife) (state 1; figs. 87, 91, 97). Apoclea algira, Philodicus tenuipes, Philonicus albiceps, Proctacanthus philadelphicus, Satanas gigas, and Zosteria rosevillensis (all Asilinae) possess spurs on the cerci that are probably used in the same way as acanthophorite spurs (state 2; fig. 95). Hull (1962) and Papavero (1973a) stated that these cercal spurs are situated on T10 or acanthophorite plates, respectively, but there is no evidence for this hypothesis as T10 is clearly differentiated from the cerci in these species (fig. 95). Among the outgroup taxa, all taxa with acanthophorite plates also possess spurs (state 1, see character 172).

175. Number of spermathecae: (0) three; (1) two; (2) one (length  =  6; CI  =  0.33; RI  =  0.20). The number of spermathecae varies little in Asilidae and with the exception of Myaptex brachyptera and Proctacanthus philadelphicus (Asilinae), which possess only two spermathecae (state 1), robber flies have three spermathecae (state 0). Among the outgroup taxa, Afroleptomydas sp. (Clanwilliam), Nemomydas brachyrhynchus, and Pseudonomoneura hirta (Mydidae), Nemestrinidae, and Hemigephyra atra (Therevidae) possess only two spermathecae (state 1) and Mitrodetus dentitarsis (Mydidae) has only a single spermatheca (state 2).

176. Development of spermathecae: (0) all equally large spermathecae; (1) median spermatheca larger than lateral ones; (2) median spermatheca smaller than lateral ones (length  =  4; CI  =  0.50; RI  =  0.00). The development of each of the spermathecal reservoirs varies little in robber flies with the exception of Leptogastrinae. Here, Euscelidia pulchra and Lasiocnemus lugens possess smaller lateral spermathecae in comparison to the median one (state 1; fig. 99) and Beameromyia bifida possesses a smaller median spermathecal reservoir (state 2). Among the outgroup taxa, Opomydas townsendi (Mydidae) also possesses reduced lateral spermathecae (state 1).

Figure 99

Female spermathecae of Lasiocnemus lugens (Leptogastrinae) in lateral view. Morphological features labeled; spermathecal reservoirs grey; second lateral spermathecal duct and reservoir, in this view situated ventral to median reservoir, omitted. Scale line  =  0.5 mm.

177. Length of common spermathecal duct: (0) short, not extending beyond tip of furca; (1) long, extending beyond tip of furca (length  =  12; CI  =  0.08; RI  =  0.47). Theodor (1976) and Yeates (1994) discussed the variation in length of the common spermathecal duct in Asilidae. Most assassin flies possess a short common spermathecal duct that does not extend beyond the tip of the furca anteriorly (state 0). The following species exhibit a long common duct: Apoclea algira (see fig. 399 in Theodor, 1976: 164), Colepia rufiventris, Neoitamus cyanurus, Neolophonotus bimaculatus, Philodicus tenuipes, Promachus amastrus, and Threnia carbonaria (all Asilinae); Brachyrhopala ruficornis, Cabasa pulchella, Codula limbipennis, and Chrysopogon pilosifacies (all Brachyrhopalinae); Laphystia species, Nusa infumata, Perasis transvaalensis, and Protometer sp. (El Tuparro) (all Laphriinae); Tipulogaster glabrata, Leptogaster species, and Lasiocnemus lugens (fig. 99) (all Leptogastrinae); Cophinopoda chinensis (Ommatiinae); and Plesiomma sp. (Guanacaste) (Stenopogoninae) (state 1; fig. 99). Among the outgroup taxa, all species possess a short common spermathecal duct (state 0).

178. Length of individual spermathecal ducts: (0) short; (1) long (length  =  9; CI  =  0.11; RI  =  0.60). The length of the individual spermathecal ducts from the common duct to the reservoir varies considerably within Asilidae. A short duct is found in Apoclea algira (see fig. 399 in Theodor, 1976: 164), Colepia rufiventris, Neolophonotus bimaculatus, Philodicus tenuipes, and Threnia carbonaria (all Asilinae); Blepharepium cajennensis, Diogmites grossus, and Neodiogmites melanogaster (all Dasypogoninae); Andrenosoma cyrtoxys, Dasyllis funebris, and Pilica formidolosa (all Laphriinae); Leptogastrinae except Acronyches maya (fig. 99); and some Mydidae species (state 0). All remaining species possess a long spermathecal duct (state 1).

179. Ejection apparatus: (0) absent; (1) present (length  =  8; CI  =  0.12; RI  =  0.22). Theodor (1976) illustrated the spermathecae of many Asilidae species in detail and discussed the development of an ejection apparatus at the base of the spermathecal duct. Within Asilinae the ejection apparatus can be absent, and this is found in the following species (state 0): Apoclea algira (see fig. 399 in Theodor, 1976: 164), Dasophrys crenulatus, Neolophonotus bimaculatus, Philodicus tenuipes, Promachus amastrus, and Zosteria rosevillensis. It has to be pointed out that it is difficult to observe the ejection apparatus with a stereomicroscope and I rely in some instances on Theodor's (1976) drawings and comments of congeneric species to interpret the presence/absence of the ejection apparatus. Among the outgroup taxa, it is absent in Mitrodetus dentitarsis, Pseudonomoneura hirta, Rhaphiomidas maehleri (Mydidae), and Nemestrinidae (state 0) and present in all other species (state 1).

180. Development of spermathecal reservoir: (0) reservoir not differentiated from spermathecal ducts; (1) reservoir formed by more or less expanded ducts to sac-shaped reservoir; (2) reservoir formed by more or less expanded and coiled ducts; (3) reservoir spherical (length  =  46; CI  =  0.06; RI  =  0.60). The shape of the spermathecal reservoir varies considerably in Asilidae and Asiloidea in general (see illustrations in Theodor, 1976; Yeates, 1994). I attempt to code this character in a meaningful and discrete way so that no overlap between the character states is possible (see fig. 100). Forty-three Asilidae species have an unmodified reservoir that is basically the same width as the spermathecal ducts (state 0). Reservoirs composed by more or less expanded ducts that might be slightly curved are found in all Leptogastrinae and Ommatiinae; many Asilinae and Dasypogoninae; and a few other species (state 1). Coiled ducts of varying degree of expansion can be found in most Laphriinae; Stenopogoninae; and a few other taxa (state 2). An entirely spherical reservoir is present in some Asilinae; Stenopogoninae; and Tillobromatinae (state 3). Among the outgroup taxa, species possess either spherical reservoirs (Apsilocephalidae, Bombylius major (Bombyliidae), Nemestrinidae, Therevidae, state 3) or an unmodified reservoir that is as wide as the spermathecal duct (most Mydidae, state 0).

Figure 100

Schematic illustration of female spermathecal reservoirs as coded in character 180. Solid lines  =  reservoir; dotted lines  =  spermathecal duct; numbers refer to character states; not to scale.

181. Sclerotization of spermathecal reservoir: (0) not sclerotized; (1) weakly sclerotized; (2) heavily sclerotized (length  =  42; CI  =  0.04; RI  =  0.42). Yeates (1994) used the sclerotization of the spermathecal reservoir as a character in the Bombyliidae phylogeny and mentioned that it varies considerably within Asilidae. In fact, all three characters states are found in Asilidae and every subfamily taxon has representatives with unsclerotized, weakly sclerotized or heavily sclerotized spermathecae. Among the outgroup taxa, all character states are found as well.

182. Length spermathecae: (0) occupying at most posterior three segments; (1) occupying more than the posterior three segments (length  =  14; CI  =  0.07; RI  =  0.51). In general, the spermathecae occupy only the posterior two abdominal segments (state 0), but in a few Asilidae species the spermathecae are longer and occupy large extents of the abdomen (state 1). These species are: several Brachyrhopalinae; some Dasypogoninae and Dioctriinae; Nusa infumata (Laphriinae); many Stenopogoninae; a few Stichopogoninae; Tillobroma punctipennis (Tillobromatinae); and all Trigonomiminae (state 1). Among the outgroup taxa, the spermathecae only occupy at most the posterior three segments (state 0).

183. Median sclerite at posterior end of furca (remnant of T9?): (0) absent; (1) present (length  =  6; CI  =  0.16; RI  =  0.79). Theodor (1976: 5; fig. 1) illustrated a median sclerite at the posterior end of the furca, but did not label or discuss it. This sclerite is easily observed when present and connects the furca to T10 (see figs. 342, 349, 399 in Theodor, 1976). It is present in Asilinae, with the exception of Blepharotes splendidissimus, Colepia rufiventris, Heligmonevra laevis, Lycomya germainii, Neoitamus cyanurus, Neolophonotus bimaculatus, Nomomyia murina, Pogonioefferia pogonias, Satanas gigas, Threnia carbonaria, and Zosteria rosevillensis; as well as all Andrenosomatini (Laphriinae) (state 1). Judging by the position of this sclerite and its connection of S9 (furca) and T10 it could be interpreted as the remnant of T9. Among the outgroup taxa, the median sclerite is always absent (state 0).

184. Development of furca (S9): (0) single sclerite; (1) divided anteriorly into two lateral sclerites (length  =  13; CI  =  0.07; RI  =  0.29). Theodor (1976) and Yeates (1994) observed that the furca in Asilidae can be composed of a single sclerite or be divided anteriorly into two sclerites. The majority of Asilidae species possesses an undivided furca (state 0; fig. 99). A divided furca is found in Dioctria species, Eudioctria albius, Echthodopa pubera, and Nannodioctria sp. (formicaphaga) (all Dioctriinae); Cerotainia albipilosa, Goneccalypsis argenteoviridis, and Loewinella nigripes (all Laphriinae); Acronyches maya, Beameromyia bifida, and Leptogaster cylindrica (see fig. 8 in Theodor, 1976: 16) (all Leptogastrinae); Lasiopogon aldrichii (Stichopogoninae); Trigonomima sp. (anamaliensis) (Trigonomiminae); and Trichoura sp. (Tierberg) (Willistonininae) (state 1). Among the outgroup taxa, Opomydas townsendi and Rhaphiomidas maehleri (Mydidae) and Hemigephyra atra (Therevidae) possess a divided furca (state 1).

185. Furcal apodeme: (0) absent; (1) present, long platelike apodeme; (2) present, short platelike apodeme; (3) present, two lateral projections forming divided apodeme (length  =  30; CI  =  0.10; RI  =  0.74). Reichardt (1929) and Theodor (1976) discussed and illustrated the furca of several species pointing out that the development of the anterior furcal apodeme (termed Gabel by Reichardt) varies within Asilidae. All Dioctriinae and Leptogastrinae; most Dasypogoninae and Laphriinae; and a few other species lack a furcal apodeme (state 0, see figs. 8, 23, 39, 184, 217 in Theodor, 1976). A long, narrow platelike apodeme is found in all Asilinae; Bathypogoninae; Andrenosomatini (Laphriinae); Cophinopoda chinensis and Ommatius species (Ommatiinae); Phellinae; Damalis species (Trigonomiminae); and a few other species (state 1, see figs. 228, 342, 349, 357, 399 in Theodor, 1976). A short, platelike apodeme is present in many Brachyrhopalinae; Stenopogoninae; Stichopogoninae; Tillobromatinae; Trigonomiminae; Willistonininae; and a few other species (state 2, see figs. 143, 154 in Theodor, 1976). A medially divided apodeme that appears bifid is present in Cabasa pulchella (Brachyrhopalinae) as well as Connomyia varipennis, Creolestes nigribarbis, and Microstylum sp. (Karkloof) (all Stenopogoninae) (state 3). Among the outgroup taxa, Apsilocephalidae possess state 2 and states 0, 1, and 3 are found in representatives of the Mydidae.

Male Abdomen

Figures 101–102 illustrate the male hypopygial features referred to in the following section. In some instances, reference is made to illustrations of species represented in the present study published by Theodor (1976) who studied the external and internal male terminalia in detail.

Figure 101

Male hypopygium and aedeagus of Lasiocnemus lugens (Leptogastrinae). Hypopygium in lateral view and aedeagus in dorsal and lateral view. Morphological features labeled; vestiture omitted; abbreviations see Materials and Methods. Scale line  =  1 mm.

Figure 102

Hypopygium of Lamyra gulo (Laphriinae) in lateral and ventral view. Morphological features labeled; vestiture omitted; abbreviations see Materials and Methods. Scale line  =  1 mm.

186. Lateral macrosetae on T2: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  10; CI  =  0.20; RI  =  0.83). See character 160 and table 4. State 1 is also found in Pegesimallus laticornis (Dasypogoninae); Daspletis stenoura (Stenopogoninae); and Oligopogon sp. (Cape Recife). Among the outgroup taxa, lateral macrosetae are never present (state 0).

187. Lateral macrosetae on T3: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  10; CI  =  0.20; RI  =  0.82). See character 160 and table 4. State 1 is also found in Ommatius tibialis (Ommatiinae); Daspletis stenoura (Stenopogoninae); and Oligopogon sp. (Cape Recife).

188. Lateral macrosetae on T4: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  12; CI  =  0.16; RI  =  0.75). See character 160 and table 4. State 1 is also found in Ommatius tibialis (Ommatiinae); Daspletis stenoura (Stenopogoninae); and Oligopogon sp. (Cape Recife).

189. Lateral macrosetae on T5: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  16; CI  =  0.12; RI  =  0.62). See character 160 and table 4. State 1 is also found in Afroestricus chiastoneurus and Ommatius tibialis (Ommatiinae); Daspletis stenoura (Stenopogoninae); and Oligopogon sp. (Cape Recife).

190. Lateral macrosetae on T6: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  15; CI  =  0.13; RI  =  0.31). See character 160 and table 4. State 1 is also found in Afroestricus chiastoneurus and Ommatius tibialis (Ommatiinae) and Oligopogon sp. (Cape Recife).

191. Lateral macrosetae on T7: (0) absent; (1) marginal (m l) setae present; (2) medial setae present (length  =  5; CI  =  0.40; RI  =  0.40). See character 160 and table 4. State 1 is also found in Afroestricus chiastoneurus and Ommatius tibialis (Ommatiinae) and Oligopogon sp. (Cape Recife).

192. Anterior margins of T3–8: (0) straight, no apodemes; (1) two lateral apodemes (length  =  1; CI  =  1.00; RI  =  1.00). Yeates and Irwin (1996) established the presence of two lateral apodemes on the anterior margin of the male T3–8 as an autapomorphy of a subordinate taxon of Mydidae. Two lateral apodemes (state 1) are found here in Mydidae species, with the exception of Neorhaphiomidas sp. (Mandurah) and Rhaphiomidas maehleri, and an entire anterior margin in all other taxa (state 0).

193. Abdominal segments: (0) T1–T8 and S1–S8 well developed; (1) T1–T7 and S1–S7 well developed, T8 + S8 reduced to ring of sclerites; (2) T1–T7 well developed, T8 entirely lost; (3) T1–T6 and S1–S6 well developed, T7–T8 and S7 reduced, S8 well developed; (4) T1–T8 and S1–S7 well developed, S8 reduced; (5) T1–T7 and S1–S8 well developed, T8 reduced to narrow sclerite; (6) T1–T6 and S1–S6 well developed, T7–T8 and S7–S8 reduced to ring of sclerites (length  =  28; CI  =  0.21; RI  =  0.68). The reduction of tergites or sternites just anterior to the male terminalia can be found in many Asilidae species. Sixty-nine out of 158 species possess tergites or sternites that are reduced at least to some extent. Hermann (1920) based his Prytaninae (unavailable name: Sabrosky, 1999; Dikow, 2004) on the reduction of the 6th abdominal segment. The different character states are found widely within Asilidae, but in all Asilinae; Bathypogoninae; Leptogastrinae; Ommatiinae; Phellinae; and Tillobromatinae; abdominal segments 1–8 are well developed (state 0). State 0 is also found in most Brachyrhopalinae; Stenopogoninae except Daspletis stenoura; and Willistonininae except Sisyrnodytes sp. (Gamka). State 1 is found in many Dasypogoninae; Lasiopogon species (Stichopogoninae); Damalis species (Trigonomiminae); and a few other species. T8 seems to be entirely lost in Rhipidocephala sp. (Harold Johnson) and Trigonomima sp. (anamaliensis) (Trigonomiminae) (state 2). The highest degree of reduction is found within Laphriinae where character states (3), (5), and (6) are exclusively found and the character state distribution for these species is summarized in table 5. State 4 is found in a few Brachyrhopalinae and Dasypogoninae; Dioctriinae except Broticosia paramonovi; Stichopogoninae except Lasiopogon species; and Holcocephala species (Trigonomiminae). Among the outgroup taxa, Poecilognathus sp. (El Hacha) (Bombyliidae) and Nemestrinidae possess state 1 whereas Mydas clavatus, Nemomydas brachyrhynchus, Neorhaphiomidas sp. (Mandurah), Opomydas townsendi, and Pseudonomoneura hirta (all Mydidae) possess state 5.

Table 5

Development of abdominal tergites and sternites of segments 6–8 in males of species of Laphriinae (character 193) Symbols: •  =  sclerites developed normally; ▵  =  sclerites reduced to narrow sclerite

194. Shape of S8: (0) simple, rectangular sclerite or reduced to narrow sclerite; (1) distal margin with protuberances; (2) distal margin with flaplike sclerite (length  =  5; CI  =  0.40; RI  =  0.00). Reichardt (1929) described and illustrated an elongated and setose S8 in Tolmerus atricapillus (as Machimus). This is in contrast to most other Asilidae species in which S8 is simple and no distal projections are developed (state 0). In the present analysis projections are found only in a few species of Asilinae. Lochmorhynchus albicans, Machimus chrysitis, and Tolmerus atricapillus possess a projection of S8 (state 1), which is well connected to the remaining part of S8. Afromochtherus mendax, Heligmonevra laevis, and Zosteria rosevillensis possess a projection on S8, but here it is not visible in dried specimens and can be examined only when the abdomen is macerated in KOH because it is joined to the sclerite by a piece of lightly sclerotized cuticle and can be flipped underneath S8 (state 2). Among the outgroup taxa, S8 is always simple (state 0). Species in which S8 is entirely lost are coded as inapplicable (-) (see character 193).

195. Rotation of hypopygium: (0) not rotated; (1) rotated by 90°–180° (length  =  18; CI  =  0.05; RI  =  0.73). The rotation of the hypopygium within Asilidae and particularly in Laphriinae, where it is most widespread, has been observed by Snodgrass (1902), Reichardt (1929), Karl (1959), and subsequent authors. Here, only the rotation of more than 90° is considered as a rotation of a few degrees may result from the preservation method of individual specimens. The direction of the rotation is not considered. Dasypogoninae, with the exception of Diogmites grossus, Saropogon dispar, and Thereutria amaraca; Laphriinae except Perasis transvaalensis; many Stenopogoninae; Tillobromatinae; Damalis species (Trigonomiminae); and a few other Asilidae species possess rotated hypopygia (state 1). Among the outgroup taxa, Bombylius major (Bombyliidae) possesses a rotated hypopygium (state 1) whereas all remaining species exhibit unrotated hypopygia (state 0).

196. Direction of hypopygium: (0) pointing posteriorly; (1) pointing dorsally (length  =  4; CI  =  0.25; RI  =  0.00). The hypopygium is generally positioned in line with the abdominal segments (state 0). In a few species of Asilinae, the hypopygium is positioned perpendicular to the abdomen and held upright (state 1; fig. 3). This was observed by Karl (1959) for Erax Scopoli, 1763 who stated that the gonocoxites are angled proximally and therefore pushing the remaining hypopygium dorsad. Species exhibiting state 1 are Clephydroneura sundaica, Eichoichemus pyrrhomystax, Neoitamus cyanurus, and Pogonioefferia pogonias (all Asilinae). Among the outgroup taxa, the hypopygium is always pointing posteriorly (state 0).

197. Development of epandrium: (0) a single sclerite, fused more or less entirely medially; (1) separated medially and joining proximally; (2) separated medially and far apart proximally (length  =  24; CI  =  0.08; RI  =  0.70). Reichardt (1929), Hardy (1948), Karl (1959), Papavero (1973a), Theodor (1976), and Sinclair et al. (1994) observed the different development of the epandrium, which may be entirely fused medially to form a single sclerite (state 0; figs. 102–103), or both epandrial halves may be separated medially (state 1; figs. 104, 113; state 2; fig. 105). All three character states are found widely within Asilidae and the outgroup taxa. A single epandrium is found in Austrosaropogon nigrinus, Brachyrhopala ruficornis, Cabasa pulchella, Chrysopogon pilosifacies, Codula limbipennis, Heteropogon manicatus, and Nannocyrtopogon lestomyiformis (all Brachyrhopalinae); Cyrtophrys attenuatus, Dasypogon diadema, Deromyia fuscipennis, Lestomyia fraudiger, Megapoda labiata, Molobratia teutonus, Pegesimallus laticornis, Pseudorus distendens, and Senobasis corsair (all Dasypogoninae); Dicolonus simplex, Dioctria species, Eudioctria albius, and Nannodioctria sp. (formicaphaga) (all Dioctriinae); Laphriinae; Connomyia varipennis (Stenopogoninae); Lissoteles aquilonius, Stichopogon species, and Townsendia albomacula (all Stichopogoninae); and Sisyrnodytes sp. (Gamka) (Willistonininae) (state 0; fig. 103). Holopogon priscus (Brachyrhopalinae); Cophinopoda chinensis, Ommatius incurvatus, and Ommatius tibialis (all Ommatiinae); and Willistonina bilineata (Willistonininae) possess a medially divided epandrium, but the epandrial halves are set far apart from each other and do not join proximally (state 2; fig. 105). The highest morphological diversity of epandrial development is found within Trigonomiminae in that Damalis annulata and Damalis sp. (Palatupana) possess a single epandrium (state 0), Rhipidocephala sp. (Harold Johnson) and Trigonomima sp. (anamaliensis) possess a medially divided epandrium (state 1), and Damalis monochaetes and Holcocephala species possess a medially divided epandrium in which the epandrial halves are set far apart (state 2; fig. 105). Among the outgroup taxa, Apioceridae, Mydidae, with the exception of Afroleptomydas sp. (Clanwilliam) and Opomydas townsendi, and Scenopinidae possess a medially divided epandrium (state 1) and all remaining species possess a single epandrial sclerite (state 0).

Figure 103–105

SEM micrographs and drawing of male hypopygia in dorsal view. 103. Hoplistomerus nobilis (Laphriinae). 104. Euscelidia pulchra (Leptogastrinae). 105. Damalis monochaetes (Trigonomiminae). Scale lines  =  0.5 mm.

198. Surstylus on epandrium: (0) absent; (1) present, distal two-thirds of epandrium form a secondary, movable lobate; (2) present, small, movable epandrial lobes connected to subepandrial sclerites (length  =  2; CI  =  1.00; RI  =  1.00). Karl (1959) and Martin (1968) discussed the division of the epandrium into a fixed, proximal sclerite and a movable, distal surstylus in many species of Leptogastrinae (sensu Martin). Surstyli are found within Eremoneura in a number of Empidoidea taxa and all Cyclorrhapha (e.g., Cumming et al., 1995; Sinclair and Cumming, 2006) and the homology of the structure found in Asilidae and in Eremoneura is not established (Griffiths, 1994). However, Emden and Hennig (1970: 341) and Torre-Bueno (Nichols, 1989: 735) define surstyli as “secondary lobate differentiations of epandria” and this accurately describes the structure found in Leptogastrinae (sensu Martin) (state 1, 101). Emden and Hennig (1970) studied Dioctria cothurnata Meigen, 1820 (Dioctriinae) and labeled a small, distal lobe of the epandrium as a surstylus. A similar, nonmovable distal projection of the epandrium is also evident in many Laphriinae, e.g., Lamyra gulo (fig. 102), and even in Mydidae (Mydas clavatus). Adisoemarto and Wood (1975) used the development of the surstylus in their phylogenetic analysis of Dioctriinae as a character, but upon close examination it is evident that the structures described are not the same that Emden and Hennig (1970) labeled as such and are rather lateral projections of the subepandrial sclerite (see character 211) and not part of the epandrium. Therefore, these structures are not considered homologous to the state found in Leptogastrinae (sensu Martin) as was also emphasized by Karl (1959). Sinclair et al. (1994) reviewed the surstylus found in Apsilocephalidae and this structure is not homologous to the surstylus found in Leptogastrinae (sensu Martin) as it is connected to the subepandrial sclerite (state 2). Ovtshinnikova (1989) established that the surstyli in Leptogastrinae (sensu Martin) are moved by the tergosternal muscle M51, which is connected to the lateral parts of the gonocoxites and the distal part of the epandrium in Asiloidea (Ovtshinnikova and Yeates, 1998). Griffiths (1994), therefore, concluded that surstyli in Leptogastrinae (sensu Martin) are not appendages ( =  styli) musculated from the epandrium itself and therefore should not be called surstyli. For lack of a better term, I use surstyli in the following sense: secondary, movable lobate differentiations of epandria (state 1; figs. 101, 104). These surstyli are found in Leptogastrinae, with the exception of Acronyches maya. Among the outgroup taxa, surstyli are absent (state 0) except for Apsilocephalidae (state 2).

199. Hypandrium shape: (0) well developed, rectangular or square sclerite; (1) well developed, triangular sclerite; (2) reduced or not differentiated (length  =  36; CI  =  0.05; RI  =  0.67). Karl (1959) used the rectangular shape and the reduction of the hypandrium as characters in his phylogenetic study on Asilidae. Papavero (1973a) discussed the shape of the hypandrium in several taxa and postulated that a free, triangular hypandrium is the plesiomorphic state in Asilidae. The different shapes are divided here into three states as these seem to be easily delineated. A well-developed, rectangular or square hypandrium is found in Asilinae except Congomochtherus inachus; Psilocurus modestus and Zabrops tagax (Laphriinae); Leptogaster arida and Tipulogaster glabrata (Leptogastrinae); Ommatiinae except Michotamia aurata; and a few species of Brachyrhopalinae; Dasypogoninae; Dioctriinae; and Stenopogoninae (state 0; figs. 106, 110, 113). A triangular hypandrium is illustrated in fig. 102 (state 1). The hypandrium can be reduced or is at least not differentiated from the gonocoxites (see character 200), which is found in many species of Laphriinae; Acronyches maya, Beameromyia bifida, and Euscelidia pulchra (all Leptogastrinae); Stichopogoninae except Lasiopogon species; Trigonomiminae except Rhipidocephala sp. (Harold Johnson); and Trichoura sp. (Tierberg) (Willistonininae) (state 2; figs. 108–109). All remaining species possess a well-developed, triangular hypandrium. Among the outgroup taxa, all three states are found.

Figure 106–111

SEM micrographs of male hypopygia in ventral view. 106. Dysmachus trigonus (Asilinae). 107. Philonicus albiceps (Asilinae), detail of gonocoxite and gonostyli. 108. Laphria flava (Laphriinae). 109. Hoplistomerus nobilis (Laphriinae). 110. Dioctria atricapillus (Dioctriinae). 111. Damalis monochaetes (Trigonomiminae). Scale lines  =  200 µm.

Figure 112–114

Hypopygia and aedeagus. 112. Stichopogon trifasciatus (Stichopogoninae), hypopygium lateral, gonocoxal apodeme grey. 113. Philodicus tenuipes (Asilinae), hypopygium lateral and dorsal, gonocoxal apodeme grey. 114. Philodicus tenuipes, aedeagus lateral, sperm sac grey. Vestiture omitted. Scale lines  =  1 mm.

200. Gonocoxites fusion when hypandrium reduced: (0) not fused medially; (1) fused proximomedially (length  =  4; CI  =  0.25; RI  =  0.72). Snodgrass (1902), Cole (1927), and Reichardt (1929) labeled the dorsal part of the rotated hypopygium of species of Laphria and Dasyllis as S9 ( =  hypandrium) and therefore postulated that the gonocoxites are absent. Karl (1959), however, showed that the dorsal structures in Laphria are in fact the gonocoxites, postulating that the hypandrium is reduced and that the gonocoxites are sometimes fused medially. This hypothesis is followed here when coding the development of the hypandrium. Thirty-five Asilidae species possess a reduced hypandrium and in most of these species the gonocoxites are not fused medially (state 0). In Andrenosomatini except Hyperechia nigripennis as well as Hoplistomerus nobilis, Laphriini, Laphystia aegyptiaca, and Stiphrolamyra angularis (all Laphriinae) the hypandrium is reduced and the gonocoxites are fused proximomedially (state 1; figs. 108–109). Among the outgroup taxa, no species with a reduced hypandrium possesses medially fused gonocoxites. All species with a well-developed hypandrium (see character 199) are coded as inapplicable (-).

201. Hypandrium and epandrium: (0) not approximating, separated by gonocoxites; (1) approximating, but not fused proximally; (2) fused at least partially; (3) fused, connected by narrow sclerites proximally (length  =  16; CI  =  0.18; RI  =  0.83). Both Yeates (1994) and Sinclair et al. (1994) independently observed the proximal connection of the epandrium and hypandrium in some species of Asilidae and postulated this character as a possible autapomorphy of robber flies. Separated epandria and hypandria are found in Ceraturgus fasciatus, Dasycyrton gibbosus, and Heteropogon manicatus (all Brachyrhopalinae); Dioctriinae; Laphriinae except Stiphrolamyra angularis; Connomyia varipennis (Stenopogoninae); and Acnephalum cylindricum and Sisyrnodytes sp. (Gamka) (Willistonininae) (state 0; fig. 102). Epandrium and hypandrium are approximating and fused at least partly in Neolophonotus bimaculatus (Asilinae); Holopogon priscus and Leptarthrus brevirostris (Brachyrhopalinae); Cyrtophrys attenuatus, Deromyia fuscipennis, Megapoda labiata, Pegesimallus laticornis, Pseudorus distendens, and Senobasis corsair (all Dasypogoninae); Beameromyia bifida and Euscelidia pulchra (Leptogastrinae); Damalis sp. (Palatupana), Holcocephala species, and Trigonomima sp. (anamaliensis) (all Trigonomiminae); and Trichoura sp. (Tierberg) (Willistonininae) (state 2; fig. 101). A proximal sclerite originating from the epandrium connecting to the hypandrium is found in Lissoteles aquilonius and Stichopogon species (Stichopogoninae) (state 3; fig. 112). In all species of Bathypogoninae; Ommatiinae; Phellinae; and Tillobromatinae; and the remaining species not mentioned above, the epandrium and hypandrium approximate, but are not fused (state 1). Among the outgroup taxa, the epandrium and hypandrium are always separated by the gonocoxites (state 0).

202. Hypandrium and gonocoxites: (0) gonocoxites entirely free; (1) gonocoxites fused to hypandrium; (2) gonocoxite-hypandrial complex (length  =  23; CI  =  0.08; RI  =  0.73). In general, the gonocoxites are free or only proximally associated with the hypandrium so that they can be moved in a vertical plane. Martin (1968) found that in most species of Leptogastrinae the gonocoxites and hypandrium are fused. A distinct suture is evident between the two parts in lateral view and this is found in all species of Asilinae; most Brachyrhopalinae; Dasypogoninae; with the exception of Saropogon dispar and Thereutria amaraca; Dioctriinae except Eudioctria albius; Hoplistomerus nobilis, Laxenecera albicincta, Nusa infumata, and Trichardis effrena (all Laphriinae); Emphysomera species and Michotamia aurata (Ommatiinae); Stenopogoninae, with the exception of Gonioscelis ventralis and Prolepsis tristis; and Tillobromatinae except Lycostommyia albifacies (state 0; figs. 101, 110, 113). The fusion of the hypandrium and gonocoxites is not uncommon among Asilidae species and is found in all Bathypogoninae; most Laphriinae; Lasiocnemus lugens, Leptogaster species, and Tipulogaster glabrata (all Leptogastrinae); Cophinopoda chinensis and Ommatius species (Ommatiinae); Phellinae; Stichopogoninae; Rhipidocephala sp. (Harold Johnson) (Trigonomiminae); and a few other species not mentioned above (state 1; figs. 101, 108–109, 112). In some species, no suture is visible and this state is here referred to as a gonocoxite-hypandrial complex, which is found in Acronyches maya, Beameromyia bifida, and Euscelidia pulchra (all Leptogastrinae); most Trigonomiminae; and Trichoura sp. (Tierberg) (Willistonininae) (state 2; fig. 111). It is important to verify whether hypandrium and gonocoxite are fused or whether one or the other is simply reduced (see character 200). Karl (1959) postulated that the hypandrium is more often reduced than the gonocoxites. The presence of gonostyli is an indicator that gonocoxite and hypandrium are fused. Among the outgroup taxa, it is assumed that the gonocoxite and hypandrium are fused into a gonocoxite-hypandrial complex in Mydidae species, with the exception of Opomydas townsendi (state 2). In Mydidae, the gonostyli are reduced (see character 205) and there is no evidence to support fusion over reduction or vice versa. Therevidae possess entirely free gonocoxites (state 0) whereas in Apioceridae, Apsilocephalidae, Bombyliidae, and Scenopinidae, the gonocoxites are fused to the hypandrium (state 1). Nemestrinidae possess a gonocoxite-hypandrial complex (state 2).

203. Epandrium and gonocoxites: (0) gonocoxites or gonocoxite-hypandrial complex entirely free; (1) gonocoxites or gonocoxite-hypandrial complex partially fused to epandrium (length  =  12; CI  =  0.08; RI  =  0.52). In a similar way to the fusion of gonocoxite and hypandrium, the epandrium can also be fused to the gonocoxite in addition to the usual proximal connection found in most species. The vertical movement of the gonocoxite is reduced due to the fusion of both sclerites. Fused epandrium and gonocoxite or a gonocoxite-hypandrial complex are found in Hoplistomerus nobilis, Lamyra gulo, Laphria species, Laxenecera albicincta, Maira aenea, Nusa infumata, Perasis transvaalensis, Smeryngolaphria numitor, Stiphrolamyra angularis, and Trichardis effrena (all Laphriinae); Leptogastrinae except Acronyches maya; Ommatius tibialis and Ommatius incurvatus (Ommatiinae); Gonioscelis ventralis (Stenopogoninae); Holcocephala species and Trigonomima sp. (anamaliensis) (Trigonomiminae); and Acnephalum cylindricum (Willistonininae) (state 1; figs. 101–102). All remaining Asilidae and outgroup taxa possess free gonocoxites (state 0; figs. 112–113).

204. Gonocoxal apodeme: (0) absent; (1) present, short, at most only slightly extending hypopygium proximally; (2) present, long, extending well into abdominal segment 8 (length  =  20; CI  =  0.10; RI  =  0.66). Sinclair et al. (1994) and Cannings (2002) discussed the different development of gonocoxal apodemes in Asiloidea and Stichopogoninae, respectively. Most species of Asilidae possess a gonocoxal apodeme (state 1; figs. 112–113) and it is absent only in Pogonioefferia pogonias (Asilinae); Austrosaropogon nigrinus, Codula limbipennis, Dasycyrton gibbosus, and Rhabdogaster pedion (all Brachyrhopalinae); Dasypogon diadema, Deromyia fuscipennis, and Pegesimallus angularis (all Dasypogoninae); Laphriinae, with the exception of Aphestia annulipes, Perasis transvaalensis, Smeryngolaphria numitor, and Stiphrolamyra angularis; Leptogastrinae except Acronyches maya; Damalis species (Trigonomiminae); and Acnephalum cylindricum (Willistonininae) (state 0). In Apoclea algira, Megaphorus pulchrus, Philodicus tenuipes, and Promachus amastrus (all Asilinae), the gonocoxal apodeme is very long and extending well into abdominal segment 8 (state 2; fig. 113). Among the outgroup taxa, a short gonocoxal apodeme is present in Apioceridae, Mydidae, with the exception of Afroleptomydas sp. (Clanwilliam), Nemomydas brachyrhynchus, and Pseudonomoneura hirta, and Therevidae (state 1), and Scenopinidae possess an elongated gonocoxal apodeme (state 2).

205. Development of gonostyli: (0) absent; (1) present (length  =  1; CI  =  1.00; RI  =  1.00). Karl (1959) found that Dissmeryngodes nigripes (Macquart, 1838) (referred to as junior synonym Atomosia dispar Walker, 1850, Laphriinae) is the only robber fly he studied that lacks gonostyli. Martin (1968) mentioned additional Atomosiini genera, i.e., Atomosia, Atomosiella Wilcox, 1937, Atoniomyia Hermann, 1912, Eumecosoma Schiner, 1866, and Cerotainia, lacking gonostyli. In the present analysis, all Asilidae possess gonostyli (state 1; figs. 101–102, 106–107, 109–110). Yeates and Irwin (1996) established the lack of gonostyli as an autapomorphy of Mydidae, which was also found by Karl (1959) in Mitrodetus dentitarsis, and all mydas flies sampled here lack gonostyli (state 0). All remaining outgroup taxa possess gonostyli (state 1).

206. Gonostyli position on gonocoxites: (0) proximal; (1) medial; (2) distal (length  =  30; CI  =  0.06; RI  =  0.50). Karl (1959) and Papavero (1973a) discussed the position of gonostyli on the gonocoxites and concluded that a distal position is plesiomorphic. The position of the gonostyli is variable within Asilidae and a distal position is most common (state 2; figs. 101–102, 109–110, 112). The gonostyli are proximal in most Asilinae, with the exception of Blepharotes splendidissimus, Heligmonevra laevis, Megaphorus pulchrus, Neoitamus cyanurus, Nomomyia murina, and Threnia carbonaria (state 0; figs. 106–107). The gonostyli are situated medially in Cyrtopogon species and Nannocyrtopogon lestomyiformis (Brachyrhopalinae); Blepharepium cajennensis, Dasypogon diadema, Megapoda labiata, and Pseudorus distendens (all Dasypogoninae); Emphysomera species (Ommatiinae); Atomosia puella, Cerotainia albipilosa, Laphria species, Laxenecera albicincta, Orthogonis scapularis, Perasis transvaalensis, and Stiphrolamyra angularis (all Laphriinae); Phellus olgae (Phellinae); Ancylorhynchus fulvicollis, Prolepsis tristis, and Stenopogon rufibarbis (Stenopogoninae); Hypenetes stigmatias and Lycostommyia albifacies (Tillobromatinae); Damalis monochaetes (Trigonomiminae); and Ablautus coquilletti and Willistonina bilineata (Willistonininae) (state 1; fig. 111). Among the outgroup taxa, Apioceridae possess gonostyli medially on the gonocoxites (state 1) and the remaining species possess distal gonostyli (state 2). Mydidae is coded as inapplicable (-) as mydas flies lack gonostyli (see character 205).

207. Lateral processes of gonostyli: (0) absent; (1) present (length  =  1; CI  =  1.00; RI  =  1.00). Martin (1968) observed lateral processes of the gonostyli in all Leptogastrinae. Later, Theodor (1976: figs. 20, 21) illustrated median processes in the species of Leptogastrinae he examined and showed a large process medially to the gonostylus that appears to be movable. The lateral processes are smaller or similar in size to the gonostyli and in my interpretation are actually situated laterally to the gonostyli as was postulated by Martin (1968). Lateral processes of gonostyli are found only to be present in Leptogastrinae except Acronyches maya (state 1; fig. 101). Among the outgroup taxa, lateral processes of gonostyli are absent (state 0) and Mydidae is coded as inapplicable (-) as all species lack gonostyli (see character 205).

208. Subepandrial sclerite shape: (0) no modifications; (1) with protuberances medially (length  =  9; CI  =  0.11; RI  =  0.11). Karl (1959) illustrated comblike setae at the posterior tip of the subepandrial sclerite in Stichopogon scaliger Loew, 1847 (Stichopogoninae). Cannings (2002) discussed the setation of the subepandrial sclerite in species of Lasiopogon (Stichopogoninae) and as it is found that within Asilidae, different protuberances and setae can be found on this sclerite (see characters 208–211). In general, robber flies possess a smooth sclerite without special protuberances (state 0), but Colepia rufiventris and Lycomya germainii (Asilinae); Psilocurus modestus (Laphriinae); Stenopogon sabaudus (Stenopogoninae); Hypenetes stigmatias and Tillobroma punctipennis (Tillobromatinae); and Holcocephala species (Trigonomiminae), possess medial protuberances on the subepandrial sclerite (state 1). Among the outgroup taxa, Bombylius major (Bombyliidae) and Neorhaphiomidas sp. (Mandurah) (Mydidae) also possess medial protuberances (state 1).

209. Subepandrial sclerite distal margin: (0) simple, straight margin; (1) distal projections; (2) short, stout macrosetae (length  =  12; CI  =  0.16; RI  =  0.52). The distal margin of the subepandrial sclerite is usually simple and straight (state 0). Distal projections are present in Afromochtherus mendax, Apoclea algira, Clephydroneura species, Colepia rufiventris, Heligmonevra laevis, Machimus occidentalis, Neoitamus cyanurus, Neomochtherus pallipes, Philodicus tenuipes, Pogonioefferia pogonias, Zosteria rosevillensis, and Threnia carbonaria (all Asilinae); Dasypogon diadema (Dasypogoninae); Aphestia annulipes and Pilica formidolosa (Laphriinae); Daspletis stenoura (Stenopogoninae); and Lycostommyia albifacies (Tillobromatinae) (state 1). Short, stout macrosetae are found in Lissoteles aquilonius and Stichopogon species (Stichopogoninae) (state 2), and have been observed before by Karl (1959) in Stichopogon scaliger. Among the outgroup taxa, only Nemestrinidae possesses distal projections (state 1).

210. Subepandrial sclerite setation: (0) absent or only a few weak setae; (1) short, stout macrosetae (length  =  1; CI  =  1.00; RI  =  1.00). Cole (1927) and Cannings (2002) found short, stout macrosetae on the subepandrial sclerite surface only in species of Lasiopogon (state 1) and this is confirmed here and no other Asilidae species possesses these setae (state 0). Among the outgroup taxa, these setae are always absent (state 0).

211. Subepandrial sclerite lateral protuberances: (0) absent; (1) present (length  =  13; CI  =  0.07; RI  =  0.42). Karl (1959) described the subepandrial sclerite (referred to as ventrale Lamelle des Analkegels) and its platelike appearance in Asilidae. Adisoemarto and Wood (1975) discussed the presence of protuberances of the subepandrial sclerite within the Dioctriinae (referred to as surstyli, see character 198). In most species of Asilidae, the lateral margin of the subepandrial sclerite is smooth without any protuberances (state 0). In Threnia carbonaria (Asilinae); Ceraturgus fasciatus and Cyrtopogon species (Brachyrhopalinae); Saropogon species (Dasypogoninae); Dioctriinae except Echthodopa pubera; Cophinopoda chinensis, Emphysomera conopsoides, Ommatius incurvatus, and Ommatius tibialis (all Ommatiinae); Creolestes nigribarbis, Microstylum sp. (Karkloof), and Prolepsis tristis (all Stenopogoninae); lateral protuberances are present (state 1). Among the outgroup taxa, Mitrodetus dentitarsis (Mydidae) also possesses lateral protuberances (state 1).

212. Fusion of cerci: (0) not fused (two distinct sclerites); (1) fused to a single sclerite (length  =  18; CI  =  0.05; RI  =  0.32). The fusion of the cerci was observed by Karl (1959) and used as a character in Stichopogoninae by Cannings (2002). The two cerci are generally distinct and only join proximally in Asilidae (state 0). In Ceraturgus fasciatus, Chrysopogon pilosifacies, Cyrtopogon rattus, Holopogon priscus, and Nannocyrtopogon lestomyiformis (all Brachyrhopalinae); Cyrtophrys attenuatus, Dasypogon diadema, Lestomyia fraudiger, Megapoda labiata, and Pegesimallus laticornis (all Dasypogoninae); Psilocurus modestus, Protometer sp. (El Tuparro), and Zabrops tagax (all Laphriinae); Ancylorhynchus fulvicollis, Creolestes nigribarbis, Ospriocerus aeacus, Prolepsis tristis, and Scylaticus costalis (all Stenopogoninae); Stichopogon species (Stichopogoninae); Tillobromatinae; Damalis sp. (Palatupana) (Trigonomiminae); and Coleomyia setigera, the two cerci are fused to a single sclerite medially (state 1). Among the outgroup taxa, the cerci are never fused (state 0).

213. Lateral ejaculatory process: (0) absent; (1) present, small triangular sclerite; (2) present, large cylindrical sclerite; (3) present, wide rectangular sclerite (length  =  33; CI  =  0.09; RI  =  0.73). Karl (1959) observed the absence of lateral ejaculatory processes in Asilinae and Ommatiinae (referred to as Querapodeme). They are found here to be absent in Asilinae; Deromyia fuscipennis, Megapoda labiata, Pegesimallus laticornis, and Pseudorus distendens (all Dasypogoninae); Echthodopa pubera (Dioctriinae); Ommatiinae; Stichopogoninae except Lasiopogon species; and Oligopogon sp. (Cape Recife) (state 1; fig. 114). When the lateral ejaculatory process is present, it can be of different shape with a small triangular sclerite most commonly represented (state 1; fig. 115). A large, cylindrical process is present in Bathypogoninae; Laphriinae, with the exception of Atomosia puella, Perasis transvaalensis, and Psilocurus modestus; Leptogastrinae; Phellus olgae (Phellinae); Lycostommyia albifacies (Tillobromatinae); Damalis monochaetes (Trigonomiminae); and a few Brachyrhopalinae; Dasypogoninae; and Stenopogoninae (state 2; fig. 101). A wide rectangular sclerite is present in Brachyrhopala ruficornis and Cabasa pulchella (Brachyrhopalinae); Aphestia annulipes, Choerades bella, Lamyra gulo, and Laphria flava (all Laphriinae); and Acnephalum cylindricum and Sisyrnodytes sp. (Gamka) (Willistonininae) (state 3). Among the outgroup taxa, states (0), (1), and (2) are present.

Figure 115–116

Aedeagi. 115. Afroholopogon peregrinus (Brachyrhopalinae), lateral and ventral. 116. Pilica formidolosa (Laphriinae), lateral and dorsal. Lateral ejaculatory process grey. Scale lines  =  0.5 mm.

214. Functional aedeagal prongs: (0) one prong; (1) two prongs; (2) three prongs (length  =  11; CI  =  0.18; RI  =  0.81). Reichardt (1929), Karl (1959), and Theodor (1976) studied the aedeagus of many Asilidae in detail and emphasized the difference in number of functional aedeagal prongs as openings of the ductus ejaculatorius. A single prong is found in most Asilidae species, i.e., all Brachyrhopalinae; Dasypogoninae; Dioctriinae; Leptogastrinae; Ommatiinae; Phellinae; Stenopogoninae; Stichopogoninae; Tillobromatinae; and Willistonininae (state 0; figs. 101, 110–111). Two prongs are found only in Myaptex brachyptera and Proctacanthus philadelphicus (Asilinae) (state 1). Three prongs are found in Asilinae, with the exception of Blepharotes splendidissimus, Colepia rufiventris, Dasophrys crenulatus, Lycomya germainii, Neoitamus cyanurus, Neolophonotus bimaculatus, Nomomyia murina, Satanas gigas, Threnia carbonaria, and Zosteria rosevillensis; Laphriinae, with the exception of Laphystia species, Perasis transvaalensis, Protometer sp. (El Tuparro), and Zabrops tagax (state 2; figs. 102, 106, 114). Among Trigonomiminae, Damalis species and Rhipidocephala sp. (Harold Johnson) possess a single prong (state 0; fig. 111) whereas Holcocephala species and Trigonomima sp. (anamaliensis) possess three prongs (state 2). Among the outgroup taxa, Afroleptomydas sp. (Clanwilliam) (Mydidae) is the only species that possesses two aedeagal prongs and all remaining species possess only a single prong (state 0).

215. Dorsal aedeagal sheath length: (0) short, sperm sac entirely free; (1) long, sperm sac entirely covered (length  =  13; CI  =  0.07; RI  =  0.85). The dorsal aedeagal sheath is usually short and does not cover the sperm sac (state 0; figs 101, 115). In Asilinae except Neoitamus cyanurus; Deromyia fuscipennis, Megapoda labiata, and Pseudorus distendens (Dasypogoninae); Echthodopa pubera (Dioctriinae); Laphriinae, with the exception of Atomosiini sensu stricto and Maira aenea; Ommatiinae; and Stichopogoninae except Lasiopogon species, the dorsal aedeagal sheath is long and covers the sperm sac entirely dorsally (state 1; fig. 114; see also figs. 229, 293, 358, 399 in Theodor, 1976). Among the outgroup taxa, Mydidae except Pseudonomoneura hirta, Nemestrinidae, Scenopinidae, and Therevidae possess a long dorsal aedeagal sheath (state 1).

216. Hypandrium posterior margin: (0) simple, no distinct projections; (1) long median projection (length  =  6; CI  =  0.16; RI  =  0.28). The distal margin of the hypandrium can be of different shape and some southern African species as well as a few other species possess a long median projection (state 1). Karl (1959) called this projection brettartiges Gebilde. Afroholopogon peregrinus (see figs. 45, 47 in Londt, 2005a: 247) and Rhabdogaster pedion (see figs. 83, 85 in Londt, 2006a: 307, Brachyrhopalinae); Broticosia paramonovi (Dioctriinae); Creolestes nigribarbis, Daspletis stenoura (see figs. 22, 24 in Londt, 1983: 296), and Microstylum sp. (Karkloof) (all Stenopogoninae); Lycostommyia albifacies (see figs. 56, 58 in Londt, 1992: 70; Tillobromatinae); and Sisyrnodytes sp. (Gamka) (Willistonininae) possess this process of unknown function (state 1) whereas all remaining Asilidae species lack a median process (state 0). Among the outgroup taxa, the posterior margin of the hypandrium is always simple (state 0).

217. Ventral median margin of dorsal aedeagal sheath: (0) weakly sclerotized, appearing incompletely closed; (1) strongly sclerotized, appearing entirely closed (length  =  12; CI  =  0.08; RI  =  0.85). Sinclair et al. (1994) discussed the incompletely closed ventral median margin of the dorsal aedeagal sheath as supporting the homology of the aedeagal sheath in all Brachycera as it is found in different taxa of Brachycera. In Asilidae, the dorsal aedeagal sheath is completely closed in Asilinae except Threnia carbonaria; Pegesimallus laticornis (Dasypogoninae); Echthodopa pubera (Dioctriinae); Laphriinae except Laphystia species; Leptogastrinae except Acronyches maya; Ommatiinae; Stichopogoninae; and Trigonomiminae, with the exception of Damalis annulata and Damalis monochaetes (state 1; figs. 114, 116). State 0 is illustrated in figure 115. Among the outgroup taxa, only Apsilocephalidae, Bombyliidae, Pseudonomoneura hirta (Mydidae), and Nemestrinidae exhibit an incompletely closed dorsal aedeagal sheath (state 0).

218. Sperm sac: (0) appearing weakly sclerotized; (1) appearing more or less heavily sclerotized (length  =  30; CI  =  0.03; RI  =  0.57). The sperm sac is always sclerotized in Asilidae, but it appears that some species have a less sclerotized sperm sac compared to other species. The majority of Asilinae; Dasypogoninae; Dioctriinae; Laphriinae; Leptogastrinae; Ommatiinae; Phellinae; and Stenopogoninae and a few Brachyrhopalinae and Willistonininae exhibit a heavily sclerotized sperm sac (state 1) whereas all remaining species exhibit a weakly sclerotized sperm sac (state 0). Among the outgroup taxa, only Apsilocephalidae, Opomydas townsendi and Pseudonomoneura hirta (Mydidae), and Nemestrinidae exhibit a weakly sclerotized sperm sac (state 0).

219. Lateral ejaculatory process and ventral aedeagal sheath: (0) lateral ejaculatory processes directed laterally or anteriorly, not surrounded by ventral aedeagal sheath; (1) lateral ejaculatory processes directed posteriorly and surrounded by ventral aedeagal sheath (length  =  6; CI  =  0.16; RI  =  0.80). When lateral ejaculatory processes are present (see character 213) they can be free or surrounded by the ventral aedeagal sheath. A free process is most commonly found within Asilidae (state 0; figs. 101, 115). Species of Laphriinae vary in their development and, with the exception of Atomosiini sensu stricto, Laphria aktis, and Perasis transvaalensis, Laphriinae exhibit state 1 (fig. 116). Among the outgroup taxa, Mitrodetus dentitarsis (Mydidae), Scenopinidae, and Therevidae possess state 1 whereas all remaining species possess state 0. All species lacking a lateral ejaculatory process are coded as inapplicable (-).

220. Shape of ejaculatory apodeme: (0) single plate of varying diameter and shape (two lateral surfaces only): (1) triangular in anterior view (two lateral and one dorsal surface); (2) diamond-shaped in anterior view (four lateral surfaces) (length  =  9; CI  =  0.22; RI  =  0.41). Reichardt (1929) and Theodor (1976) illustrated the musculature of the ejaculatory apodeme connecting to the dorsal aedeagal sheath. The apodeme itself can be of different shape and is most commonly a single plate of varying diameter, which has two lateral surfaces and is positioned vertically in the hypopygium or abdominal segment 8 (state 0; figs. 101, 114–Figure 115–116117). A triangular apodeme with two lateral and one dorsal surface is found in Blepharepium cajennensis (Dasypogoninae); Phellinae; Stichopogoninae except Lasiopogon cinctus; and Willistonina bilineata (Willistonininae) (state 1; fig. 117). A diamond-shaped apodeme with four lateral surfaces is found in Megapoda labiata (Dasypogoninae) and Nusa infumata and Smeryngolaphria numitor (Laphriinae) (state 2; fig. 117). Among the outgroup taxa, most species possess a single plate (state 0) and only Nemestrinidae possess a triangular ejaculatory apodeme (state 1).

Figure 117

Schematic drawing of ejaculatory apodemes in posterior view as coded in character 220 with character states indicated. Not to scale.

Characters Not Considered

Many characters and character complexes show a high degree of morphological variation within Asilidae, such as pruinosity on the thorax and abdomen, setal arrangement on legs, wing microtrichia covering, shape and appendages of the aedeagus, and shape of gonostyli in males. Such characters are not included here as their development is often associated with species-specific differences that are very difficult to homologize across taxa. Only features for which character states are consistently applicable across the entire Asilidae are included.

Fisher (1986) postulated a field of sensilla situated laterally on the cerci in female Andrenosomatini as an autapomorphy for this taxon. This character is not included here as it would have been necessary to take scanning electron micrographs of the ovipositor of every female specimen to identify it correctly. The value of this character is not in doubt, however. Fisher (personal commun.) studied a number of Asilinae species with respect to the presence of a crop (absent in Asilinae but present in Apocleinae sensu Papavero (1973a) and Geller-Grimm (2004)). As this character is observable only in freshly killed or specially preserved specimens, it could not be included here as the vast majority of specimens are dry-mounted, many of them considerably old.

The musculature of the male terminalia has been shown to contain characters that can be useful in delimiting taxa (Ovtshinnikova, 1989; Ovtshinnikova and Yeates, 1998), e.g., support for a clade Apioceridae, Asilidae, and Mydidae is provided by the presence of muscle M33 in these three taxa. Despite the usefulness of this character complex, no musculature is examined due to the different preservation conditions of the many specimens examined and the exemplar approach employed here.

Theodor (1976) and Yeates (1994) discussed a number of characters of the female spermathecae that are only observable under high magnification in a compound microscope, e.g., sclerotized tubules on spermathecal duct, and these characters are not considered as no slide mounts of spermathecae are prepared.

A Revised, Higher-Level Classification of Asilidae

Since the first theoretical contributions by Hennig (1950, 1966), only monophyletic taxa, i.e., taxa including all descendents, extant and extinct, of their most recent common ancestor, are recognized in phylogenetic systematics. Evolutionary taxonomists, in contrast, use anagenesis in addition to cladogenesis in deriving classifications. The result is the delimitation of taxa that do not entail every descendent of the most recent common ancestor (e.g., Mayr, 1974, 1981). It is not evident from the introduction to the two most comprehensive classifications on Asilidae by Hull (1962) and Papavero (1973a) whether they propose a cladistic or evolutionary classification. Both studies, as outlined in the introduction, are not cladistic in nature, although both authors labeled their diagrams of relationships as phylogenetic, i.e., “A provisional philogenetic [sic] arrangement of the Asilidae” (Hull, 1962: text-fig. 35) and “Hypothetical phylogeny of the subfamilies of Asilidae” (Papavero, 1973a: fig. 1). Both authors first postulated a classification from which they then deduced a phylogeny. In cladistics, by contrast, the phylogeny is postulated first, and only then is a phylogenetic classification derived, so that only monophyletic taxa are recognized. Griffiths (1974), Hennig (1975), Farris (1979), Gaffney (1979), and many later authors emphasized that “the most important biological reference system is a phylogeny and that a classification that mirrors a phylogenetic hypothesis (however transitory) is the most useful for systematists and nonsystematists” (Gaffney, 1979: 104). Wheeler (1995: 31) stated,

The present study is the most comprehensive phylogenetic analysis of Asilidae to date and there is no other choice than to also use the phylogenetic branching pattern to establish a new, monophyletic, revised classification. Admittedly, translating a phylogeny into a monophyletic classification is straightforward, but it is certainly subjective in the sense that one cannot (1) recognize every monophyletic taxon with a name and (2) objectively decide whether to call a certain taxon a tribal taxon or subfamily taxon. The only objective consideration is that when one recognizes a group of species as a taxon, one also has to recognize its sister taxon at the same rank no matter its position in the cladogram.

Unfortunately, at this stage it is impossible to write a useful dichotomous or matrix-based identification key that would place all genera of Asilidae into subordinate taxa. The desire of dipterists to have this tool is understandable, but since only 140 out of 537 genera are represented in the present analysis, no meaningful key can be written. However, table 7 summarizes the character distribution of 24 easily observed characters that can aid in placing yet unstudied genera and species. The provided diagnoses of all taxa including autapomorphies and additional apomorphic character states, which are also found in other phylogenetically unrelated robber flies, should also be consulted. Table 3 indicates the respective lengths of the proximal tarsomeres that appear to be diagnostic for at least some taxa, and so should be consulted as well.

The following discussion, also summarized in figures 120–121 and table 6, provides an overview of the taxa (subfamily and tribe) recognized in the revised classification. A numbering system is employed to indicate the subordination of taxa and the following information for each taxon is provided: autapomorphies, additional apomorphic character states that are diagnostic, and remarks with respect to the use of the family-group name by previous authors. One of the 720 most parsimonious cladograms is selected to illustrate the character optimization and character distribution (figs. 122–Figure 123Figure 124Figure 125126) although the delimitation of taxa is based on the strict consensus topology (fig. 119) as it is the summary of all equally parsimonious hypotheses. Taxa that only contain a single genus are indicated as well, although the diagnosis is based only on the single included species.

Figure 120

Summary of relationships of subfamily taxa based on strict consensus topology of figure 119 and revised classification of table 6. Character state optimization and distribution for higher-level clades mapped onto cladogram with character number above circle and character state below circle; filled circle  =  autapomorphy; open circle  =  reversal or homoplasy; clade names above branches; Bremer support/relative Bremer support below branches.

Figure 121

Summary of relationships of tribal taxa based on strict consensus topology of figure 119 and phylogenetic placement of the two oldest fossil Asilidae. Clade names as in figure 120.

Figure 122

Clade Laphriinae of figure 120 with character state optimization and distribution mapped onto cladogram. Character number above circle and character state below circle; filled circle  =  autapomorphy; open circle  =  reversal or homoplasy.

Figure 123

Clade B (Asilinae + Ommatiinae) of figure 120 with character state optimization and distribution mapped onto cladogram. Character number above circle and character state below circle; filled circle  =  autapomorphy; open circle  =  reversal or homoplasy.

Figure 124

Clade C without clade H (Bathypogoninae, Phellinae, Tillobromatinae, Dasypogoninae, Stenopogoninae, and Coleomyia setigera) of figure 120 with character state optimization and distribution mapped onto cladogram. Character number above circle and character state below circle; filled circle  =  autapomorphy; open circle  =  reversal or homoplasy.

Figure 125

Clade H without clade M (Willistonininae, Dioctriinae, Leptogastrinae, Trigonomiminae, and Oligopogon sp. [Cape Recife]) of figure 120 with character state optimization and distribution mapped onto cladogram. Character number above circle and character state below circle; filled circle  =  autapomorphy; open circle  =  reversal or homoplasy.

Figure 126

Clade M (Brachyrhopalinae and Stichopogoninae) of figure 120 with character state optimization and distribution mapped onto cladogram. Character number above circle and character state below circle; filled circle  =  autapomorphy; open circle  =  reversal or homoplasy.

Table 6

Revised classification of Asilidae and list of recognized family-group taxa. Placement of genera in each taxon is based on the present phylogenetic hypothesis

Table 7

Comparative identification matrix for the 14 subfamily taxa based on the present phylogenetic hypothesis Symbols: •  =  present in all species; ▵  =  present in majority of species; #  =  present in a few species; –  =  absent. See also table 3 for respective lengths of proximal tarsomeres

This newly revised phylogenetic classification recognizes 14 subfamilial taxa based on already published family-group names summarized in Sabrosky (1999) and Dikow (2004). The status of some of the taxa is revised, but no new family-group names are described. It should be highlighted that a number of higher-level taxa actually contain the same genera as postulated by previous authors; for example, the present delimitation of Laphriinae comprises all genera that Papavero (1973a) and Geller-Grimm (2004) placed in Laphriinae and Laphystiinae. The phylogenetic hypothesis could also have been divided into three higher-level taxa, i.e., Asilinae (including Ommatiinae,  =  taxon B), Laphriinae, and Dasypogoninae ( =  taxon C), which would be the same arrangement as postulated by Macquart in 1838. However, I did not choose to propose this classification as taxon C would be (1) morphologically very heterogeneous, (2) the large leftover group that is not easily diagnosable, and (3) the majority of generic and species diversity that, in my view, would not be investigated as thoroughly as the easily delimited Asilinae or Laphriinae.

1. Laphriinae

Laphriinae Macquart, 1838. Type genus Laphria Meigen, 1803. Species represented: 29. Figures 120–Figure 121122.

Autapomorphies: None. Additional apomorphic character states: Dorsal margin of postocciput with triangular projection (42; fig. 50), apical “seta-like” sensory element situated apically in cavity on stylus (57; fig. 57), dorsal margin of prosternum with distinct flangelike projection (75; fig. 66), metkatepisternum large and visible between mesothoracic and metathoracic coxae (ventral view) (100), setae present on lateral metathoracic coxae (112), cell r₁ closed, R₁ and R₂₊₃ meet at costa (131; figs. 73, 76), R₂₊₃ apex arching sharply anteriorly in 90° angle (142; fig. 73), R₄ strongly sinuate (144; fig. 73), female T10 lost or fused to cerci (172), male abdominal segments 1–6 well developed, T7–8 and S7 reduced while S8 is well developed (193). Laphriinae is the oldest available family-group name within this clade and includes all genera, which had been previously assigned to Laphriinae and Laphystiinae by Papavero (1973a) and Geller-Grimm (2004).

1.1. Laphystiini Hendel, 1936, new status.

Type genus: Laphystia Loew, 1847. Genera included: Laphystia. Autapomorphies: None. Additional apomorphic character states: Stylus of antenna composed of two elements (58), female common spermathecal duct long and extending beyond tip of furca (177), male ventral median margin of dorsal aedeagal sheath weakly sclerotized (217), male sperm sac appearing weakly sclerotized (218). The taxon Laphystiinae sensu Papavero (1973a) is not recovered as monophyletic in the present analysis and only the two included species of Laphystia form a clade that is here recognized as a tribal taxon.

1.2. Atomosiini Lynch Arribalzaga, 1882.

Type genus Atomosia Macquart, 1838. Genera included: Aphestia, Atomosia, Cerotainia, Goneccalypsis, Loewinella, Orthogonis, Smeryngolaphria. Autapomorphies: None. Additional apomorphic character states: Postpedicel cylindrical (same diameter throughout) (54), apical “seta-like” sensory element positioned subapically laterally on postpedicel (57), fringe of parallel setae present ventrally on metathoracic femora and tibiae (118), metathoracic tibiae straight (not arched) (119). The Atomosiini is considered here to contain also the Dasytrichini, i.e., Orthogonis and Smeryngolaphria, which is not recovered as monophyletic. The diagnostic characters mentioned above are sufficient to recognize Atomosiini and the fact that Orthogonis scapularis shares an autapomorphy with the Atomosiini sensu previous authors exemplifies their close phylogenetic relationship. Atomosiini except Smeryngolaphria are characterized by an autapomorphy, i.e., cell d is closed at the distal end only by the base of M₂ and crossvein m-m is absent or at least highly reduced (137), and additional apomorphic character states (14: 0, 40: 2, 86: 0, 203: 0). The Atomosiini sensu previous authors are characterized by an autapomorphy, i.e., anatergal setae are formed by short and stout macrosetae (82), and additional apomorphic character states (23: 2, 25: 1, 95: 1, 100: 0, 102: 3, 109: 1, 141: 0, 151: 1, 170: 0, 193: 1, 215: 0, 219: 0). The clade Goneccalypsis + Loewinella is characterized by three autapomorphies, i.e., two short, stout, erect macrosetae are present on the anterior mesonotum (87), females and males possess medial setae on T7 (165, 191), and additional apomorphic character states (50: 1, 70: 0, 130: 2, 184: 1). The clade Aphestia + (Atomosia + Cerotainia) possesses two autapomorphies, i.e., costa terminates at M₂ (140; fig. 77), females with segment 7 distinctly shortened (166), and additional apomorphic character states (167: 1, 199: 1). Aphestia is characterized by a single autapomorphy, i.e., fringe of parallel setae is only present on metathoracic tibiae (118). The clade Atomosia + Cerotainia possesses two autapomorphies, i.e., T6 is cup-shaped and the last visible tergite conceals the remaining tergites (158), male T7–8 and S7–8 are reduced to ring of sclerites (193), and additional apomorphic character states (83: 1, 130: 1, 162: 0, 163: 0, 206: 1).

1.3. Hoplistomerini Enderlein, 1936.

Type genus Hoplistomerus Macquart, 1838. Genera included: Hoplistomerus, Trichardis. Autapomorphies: None. Additional apomorphic character states: Macrosetae and setae on the postpronotum (73), postsutural dorsocentral setae absent (94), scutal setae possess a large socket (95), short and stout cone-shaped macrosetae present on metathoracic femora (116), R₂₊₃ apex arching sharply anteriorly in 90° angle (142; fig. 73), abdominal tergites indented and setae possess large sockets (151). Trichardis is assigned to this taxon for the first time.

1.4. Laphriini Macquart, 1838.

Genera included: Choerades, Lampria, Laphria, Maira. Autapomorphies: Face with circular and dorsoventrally flattened setae laterally in males and females (45), medial setae on posterior scutum present, appressed to surface and directed posteriorly (96). Additional apomorphic character states: Prementum laterally compressed proximally (18; fig. 32), pharyngeal pump with entirely sclerotized ring (40; fig. 48), antennal insertion elevated above eye margin with distinct protuberance (53). In addition to the genera listed above, Laxenecera and Nusa were regarded to belong to Laphriini by previous authors. These two genera do not group with other genera of Laphriini and render it non-monophyletic. Therefore, they are excluded from the easily diagnosed Laphriini with dorsoventrally flattened setae on the face and appressed setae on the posterior scutum.

1.5. Ctenotini Hull, 1962.

Type genus Ctenota Loew, 1873. Genera included: Lamyra, Stiphrolamyra. Autapomorphies: Posterior margin of postgenae with large flangelike projection (47). Additional apomorphic character states: Dorso-posterior margin of cibarium simple, no particular ridge connecting cornua developed (35), presutural dorsocentral setae absent (92), macrosetae and setae present on median metathoracic trochanter (114), metathoracic tibiae straight (119), tip of claws blunt (126), male abdominal segments 1–6 well developed, T7–8 and S7 reduced while S8 is well developed (193). Ctenotini is recovered as monophyletic and contains the same taxa as previously postulated.

1.6. Andrenosomatini Hull, 1962.

Type genus Andrenosoma Rondani, 1856. Genera included: Andrenosoma, Dasyllis, Hyperechia, Pilica, Proagonistes. Autapomorphies: Female T10 is divided into two heavily sclerotized sclerites that are far apart (172; fig. 88). Additional apomorphic character states: Distal maxillary palpomere laterally compressed and leaflike (13), prementum of labium dorsoventrally flattened (18), apex of labella of labium pointed (24), female with median sclerite at posterior end of furca (183), female with long platelike apodeme on furca (185), male with epandrium and gonocoxites entirely free (203). Andrenosomatini is recovered as monophyletic and contains the same taxa as previously postulated. The relationship among the genera, however, is not resolved. Fisher (1986) presented a comprehensive phylogenetic study of Andrenosomatini.

Unplaced genera: Laxenecera, Nusa, Perasis, Protometer, Psilocurus, Zabrops. These genera cannot be assigned to any available family-group taxon and their phylogenetic relationship must be evaluated using a denser taxon sampling within Laphriinae.

Non-monophyletic taxa: Dasytrichini Lynch Arribalzaga, 1882. Representatives of this taxon are now included in Atomosiini with which they have been previously associated by Hermann (1912).

Phylogenetic relationships among Laphriinae: The genera Perasis, Zabrops, Protometer, Psilocurus, Laphystia, Hoplistomerus, and Trichardis, which have been previously assigned to the Laphystiinae, form a grade of less inclusive sister taxa that are more closely related to the Laphriinae sensu stricto. Perasis transvaalensis is the sister taxon to the remaining Laphriinae, which possess two autapomorphies, i.e., females and males with medial setae on T2 (160, 186), and additional apomorphic character states (54: 2, 69: 1, 180: 2, 195: 1, 204: 0, 212: 1, 213: 2, 219: 1). Laphriinae except Perasis transvaalensis and Zabrops tagax possess four autapomorphies, i.e., females and males with medial setae on T3–4 (161, 162, 187, 188), and additional apomorphic character states (62: 1, 102: 1, 161: 2, 164: 2). The four speciose taxa Atomosiini, Laphriini, Ctenotini, and Andrenosomatini form a clade characterized by these apomorphic character states: 4: 2, 24: 2, 88: 1, 134: 0, 140: 0, 202: 1. Laphriini, Ctenotini, and Andrenosomatini form a clade (21: 1, 23: 2, 25: 1, 200: 1) and lastly Ctenotini and Andrenosomatini are sister taxa delimited by the following apomorphic character states: 28: 1, 49: 1.

2. Ommatiinae

Ommatiinae 45Hardy, 1927. Type genus Ommatius Wiedemann, 1821. Species represented: 7. Figures 120–121, 123.

Autapomorphies: Setae on anterior surface of antennal stylus (59). Additional apomorphic character states: At least some median ommatidia larger than surrounding ones (46), postpedicel short and medially broadest (54; fig. 52), postmetacoxal area entirely sclerotized (102). The Ommatiinae has been considered to be a monophyletic taxon that is usually associated with the Asilinae or Apocleinae sensu previous authors (e.g., Karl, 1959; Hull, 1962; Lehr, 1969, 1996; Papavero, 1973a; Bybee et al., 2004). In the present analysis, the included genera Afroestricus, Cophinopoda, Emphysomera, Michotamia, and Ommatius form a polytomy in the strict consensus hypothesis (fig. 119) and the genus Ommatius is paraphyletic with respect to Afroestricus chiastoneurus. Afroestricus has only recently been described for 20 Afrotropical species previously assigned to Ommatius by Scarbrough (2005) and Afroestricus chiastoneurus is the sister group to Ommatius tibialis (Nearctic), but phylogenetically unrelated to Ommatius incurvatus (Neotropical). The genus Emphysomera is monophyletic and characterized by a number of apomorphic character states (7: 3, 67: 1, 74: 1, 92: 1, 121: 0, 206: 1). The inclusion of more Ommatius species in future studies will help to clarify the monophyly of this worldwide, speciose genus.

3. Asilinae

Asilinae Latreille, 1802. Type genus Asilus Linnaeus, 1758. Species represented: 32. Figures 120–121, 123.

Autapomorphies: None. Additional apomorphic character states: Macrosetae on lateral margin of frons present (49), prosternum and proepisternum separated and prosternum triangular and pointed dorsally (74; fig. 68), female S8 proximally platelike and distally hypogynial valves forming a keel (170; figs. 90, 95), male gonocoxites entirely free (202; fig. 113). The strict consensus topology in figure 119 illustrates the unresolved relationships among Asilinae genera including the Apocleinae sensu previous authors. This highlights the fact that the characters employed here cannot group the morphologically very similar species into less inclusive taxa. The addition of additional features, additional taxa, and other data, for example DNA sequence data, will be necessary to postulate a well-corroborated phylogenetic hypothesis for Asilinae species. The monophyly of Asilinae is, however, corroborated and has never been questioned before by previous authors.

3.1. Asilini Latreille, 1802.

Genera included: Asilus. Autapomorphies: None. Additional apomorphic character states: Maxillary stipites fused entirely medially (14; fig. 34), female T2–3 without marginal setae (160, 161), male T2–5 without marginal setae (186, 187, 188, 189). In the present analysis, only the single genus Asilus is placed in the Asilini as Satanas does not group with Asilus.

3.2. Apocleini Papavero, 1973a new status.

Type genus Apoclea Macquart, 1838. Genera included: Apoclea, Megaphorus, Philodicus, Promachus. Autapomorphies: None. Additional apomorphic character states: Facial swelling extending over lower facial half (4), postmetacoxal bridge partly developed laterally, but membraneous medially (102), stump vein on R₄ present and reaching base of R₂₊₃ (146; figs. 74, 80), female with median sclerite at posterior end of furca (183), male gonocoxal apodeme long and extending well into abdominal segment 8 (204; fig. 113). The Apocleini is the remnant of the Apocleinae sensu Artigas and Papavero (1988) and Geller-Grimm (2004). However, Artigas and Papavero (1997) combined the Asilinae and Apocleinae and the placement of Apocleini within Asilinae is corroborated here. Megaphorus is the sister taxon to the remaining Apocleini, which are characterized by the following apomorphic character states: 7: 1, 85: 2, 177: 1, 179: 0, 186: 1, 187: 1, 188: 1, 189: 1. Apoclea and Philodicus form a clade supported by these apomorphic character states: 92: 0, 168: 2, 174: 2, 178: 0, 209: 1, 218: 0.

3.3. Lycomyini Lynch Arribalzaga, 1882.

Type genus Lycomya Bigot, 1857. Genera included: Lycomya. Autapomorphies: None. Additional apomorphic character states: Facial swelling extending over lower half of face (4), prementum of labium dorsoventrally flattened (18), stylus composed of two elements (58), anatergal setae present (82), postmetacoxal bridge partly developed laterally, but membraneous medially (102), setiform empodium reduced in length (127), microtrichia on posterior margin of wing arranged in two divergent planes (141), T2 longer than wide, but less than twice as long (153), male subepandrial sclerite with protuberances medially (208). Lycomyini has been described for the single Chilean genus Lycomya by Lynch Arribalzaga (1882).

3.4. Machimini Lehr, 1996.

Type genus Machimus Loew, 1849. Genera included: Machimus. This taxon is apparently non-monophyletic as the two included species of Machimus do not form a clade. In light of the unresolved relationships within the Asilinae, no comments on the monophyly of the genus or the taxon can be made.

3.5. Neomochtherini Lehr, 1996.

Type genus Neomochtherus Osten-Sacken, 1887. Genera included: Afromochtherus, Neomochtherus. Autapomorphies: None. Additional apomorphic character states: Setae on posterior basalare absent (81), female T8 with posteriorly directed setae (168). Neomochtherini has been described by Lehr (1996) to also include Dysmachus, but this genus is part of the Asilinae genera that cannot be placed in any higher-level taxon.

3.6. Philonicini Lehr, 1996.

Type genus Philonicus Loew, 1849. Genera included: Philonicus. Autapomorphies: None. Additional apomorphic character states: Microtrichia on posterior margin of wing arranged in two divergent planes (141), female T5 without any marginal setae (163), female with spurs on cerci (174; fig. 95). Philonicini has been described by Lehr (1996) based on the single genus Philonicus.

Unplaced genera: Blepharotes, Clephydroneura, Colepia, Congomochtherus, Dasophrys, Dysmachus, Eichoichemus, Heligmonevra, Lochmorhynchus, Machimus, Myaptex, Neoitamus, Neolophonotus, Nomomyia, Pogonioefferia, Proctacanthus, Satanas, Threnia, Tolmerus, Zosteria. Although Colepia, Neoitamus, and Threnia form a monophyletic clade characterized by apomorphic character states—macrosetae on lateral margin of frons (dorsal to antennal insertion) absent (49) and female with spermathecal duct long and extending beyond the tip of furca (177)—no new name is proposed because of the poor resolution of relationships within Asilinae genera and the possibility that this group might not be monophyletic once new characters or taxa are added.

4. Bathypogoninae, new status

Bathypogoninae Artigas and Papavero, 1991. Type genus Bathypogon Loew, 1851. Species represented: 1. Figures 120–121, 124.

Autapomorphies: None. Additional apomorphic character states: Macrosetae on lateral margin of frons present (49), anterior anepisternal setae absent (78), protuberance on anterior metathoracic coxae present and blunt distally (113), R₂₊₃ arching anteriorly in about 90° angle (142), R₄ strongly sinuate (144), male with large cylindrical lateral ejaculatory process (213). The Bathypogonini has been described by Artigas and Papavero (1991) as a taxon within their Stenopogoninae based on the Australian genus Bathypogon and on Carebaricus Artigas and Papavero, 1991, from Chile and Argentina. The single included species from Australia representing this taxon, Bathypogon nigrinus, does not group with any other Stenopogoninae sensu previous authors and therefore is recognized here as a separate taxon. Bathypogoninae is the sister group to the large clade D, which highlights its morphological distinctness. Carebaricus is not included in the present analysis and it cannot be confirmed whether it groups with Bathypogon. Extant representatives of Bathypogoninae are known only from Argentina, Australia, and Chile (Hennig, 1960; Artigas and Papavero, 1991).

5. Phellinae, new status

Phellinae Hardy, 1926. Type genus Phellus Walker, 1851. Species represented: 2. Figures 120–121, 124.

Autapomorphies: None. Additional apomorphic character states: Maxillae (in lateral view) proximally high and distal tip narrower (28; fig. 38), ocellar setae formed by regular setae (no macrosetae) (63), postocular setae formed by regular setae (no macrosetae) (66), three or more notopleural setae present (88), stump vein on R₄ (not reaching R₂₊₃) (146; fig. 79), M₂ and m-m aligned in a line from distal to proximal (nearly parallel to posterior wing margin) (150; fig. 81), female with 7th and following segments comprising ovipositor (167), female S8 of ovipositor keellike throughout (170; fig. 90), male ejaculatory apodeme triangular in anterior view (two lateral and one dorsal surface) (220; fig. 117). The taxon Phellini was described by Hardy (1926) to include the three Australian genera Bathypogon, Phellus, and Psilozona Ricardo, 1912. Hull (1962) transferred the Chilean genus Obelophorus into Phellini and Artigas and Papavero (1991) transferred Bathypogon to the newly described Bathypogonini. The two included genera Obelophorus and Phellus form a clade supported by a number of apomorphic character states and are recognized here as a separate taxon that is the sister taxon to clade E. This taxon is biogeographically interesting as the extant representatives are found only in Australia and Chile.

6. Tillobromatinae, new status

Tillobromatinae Artigas and Papavero, 1991. Type genus Tillobroma (Hull, 1962). Species represented: 3. Figures 120–121, 124.

Autapomorphies: None. Additional apomorphic character states: Maxillary stipites fused medially, but with V-shaped indentation (14), female spermathecal reservoir formed by more or less expanded and coiled ducts (180). The taxon Tillobromatini was described by Artigas and Papavero (1991) for the distinct New World genera Coleomyia, Euthrixius Artigas, 1971, Grajahua Artigas and Papavero, 1991, Scylaticina Artigas and Papavero, 1991, Scylaticodes Artigas and Papavero, 1991, Tillobroma, and Zabrotica Hull, 1958. Tillobroma was described by Hull (1962) as a subgenus of the South African genus Hypenetes highlighting the morphological similarity between these flies from different continents. Artigas et al. (2005) based a divergence time estimate of 180 Million years on the assumption of a sister-group relationship between Hypenetes and Tillobroma, which is not supported by the present analysis. Hypenetes and another South African genus, Lycostommyia, form a clade supported by a number of apomorphic character states (23: 2, 46: 0, 74: 2, 121: 0, 206: 1) and these two genera together form the sister group to Tillobroma. The western North American genus Coleomyia is not recovered within Tillobromatini and is here placed as incertae sedis (see below). As the genera Hypenetes, Lycostommyia, and Tillobroma form a clade distinct from other Dasypogoninae and Stenopogoninae, they are recognized here as a separate taxon. This taxon is biogeographically interesting, as the extant representatives are restricted to southern Africa and South America.

7. Dasypogoninae

Dasypogoninae Macquart, 1838. Type genus Dasypogon Meigen, 1803. Species represented: 16. Figures 120–121, 124.

Autapomorphies: None. Additional apomorphic character states: Maxillae proximally high and distal tip distinctly narrower (28; fig. 38), pharyngeal pump laterally with sclerotized ring, but medio-anteriorly unsclerotized (40; fig. 49), setae on dorsal postpedicel present (55), more than two macrosetae on median occipital sclerite present (65), large spine originating anteroventrally on prothoracic tibiae (106; figs. 70–71). The Dasypogoninae sensu previous authors has been suspected to be non-monophyletic. However, the Dasypogon-like robber flies recognized here within this taxon represent genera that have always been believed to form a monophyletic clade. The large spine on the prothoracic tibiae has been used as a diagnostic character for this group (e.g., Hull, 1962; Papavero, 1973a). Based on the phylogenetic hypothesis derived here this spine has been independently derived at least once in the Australian Brachyrhopalini and Chrysopogonini (Brachyrhopalinae). The spine has been reduced at least once within Dasypogoninae in Archilestris magnificus. Only Thereutria possesses a small S-shaped spur on the prothoracic tibiae, which is independently derived in Cophura, Leptarthrus, and Nicocles (all Brachyrhopalinae). Eleven subordinate taxa have been proposed to divide the Dasypogoninae by previous authors and with the exception of Archilaphriinae, all are included in the present analysis although some of them are only represented by a single species. Of these available and valid family-group taxa, eight are here recognized within Dasypogoninae (see below) and Brachyrhopalini and Chrysopogonini are transferred to the newly erected Brachyrhopalinae.

7.1. Dasypogonini Macquart, 1838.

Genera included: Dasypogon. Autapomorphies: None. Apomorphic character states: All ommatidia of same size (46), postpedicel cylindrical with same diameter throughout (54), regular setae form postpronotal setae (no macrosetae) (73), medial setae on posterior scutum absent (96), cell m₃ closed (135; fig. 82), cell cup closed (136; fig. 82), microtrichia on posterior wing margin arranged in two divergent planes (141), female anterior T8 simple, no apodeme present (169), female S8 platelike with hypogynial valves separated and surrounded by membrane (170), male gonocoxal apodeme absent (204), male gonostyli positioned medially on gonocoxites (206), male subepandrial sclerite with distal projections (209). Of the included species here, Lestomyia fraudiger and Austrosaropogon nigrinus are classified as Dasypogonini by Hull (1962) and Papavero (1973b). Lestomyia groups with other Dasypogoninae, but not with Dasypogon itself and is excluded here from this taxon. The Australian genus Austrosaropogon groups with other Australian genera in Brachyrhopalinae (see below).

7.2. Saropogonini Hardy, 1926.

Type genus Saropogon Loew, 1847. Genera included: Saropogon. Autapomorphies: None. Apomorphic character states: Macrosetae on dorsal frons (dorsal to antennal insertion) present (49), female spermathecal reservoir formed by more or less expanded and coiled ducts (180), male subepandrial plate with lateral protuberances (211). The Saropogonini was originally described by Hardy (1926) for the Australian genera Rachiopogon Ricardo, 1912, Erythropogon White, 1914, Questopogon Dakin and Fordham, 1922, and Neosaropogon Ricardo, 1912, and the more widespread genera Saropogon, Stenopogon, and Clinopogon Bezzi, 1910. Oldroyd (1974a, 1974b) used the taxon for a large number of genera, but it is restricted here to the single genus Saropogon.

7.3. Thereutriini Hull, 1962.

Type genus Thereutria Loew, 1851. Genera included: Thereutria. Autapomorphies: None. Apomorphic character states: Ocellar setae formed by regular setae (no macrosetae) (62), two notopleural setae present (88), presutural dorsocentral setae present (92), prothoracic tibiae with small S-shaped spur, which originates on posteroventral surface (106), regular setae on median metathoracic trochanter (no macrosetae) (114), cell r₁ closed, R₁ and R₂₊₃ as stalk vein present (131), female S8 platelike and hypogynial valves separated and surrounded by membrane (170), male hypopygium not rotated (195), male gonocoxites and hypandrium fused (202). The Thereutriini was described by Hull (1962) to accommodate the Australian genera Metalaphria Ricardo, 1912, and Thereutria.

7.4. Lastauracini Papavero, 1973b.

Type genus Lastaurax Carrera, 1949. Genera included: Neodiogmites. Autapomorphies: None. Apomorphic character states: Presutural dorsocentral setae present (92). The Lastauracini was described by Papavero (1973b) for 10 Neotropical genera and Diogmites, which is distributed in the Nearctic and Neotropical regions. The original type genus is Lastaurax Carrera, 1949, but this genus has been synonymized with Neodiogmites by Artigas and Papavero (1988), which is now the type genus of this family group. Diogmites does not group with Neodiogmites in the present analysis and therefore the Lastauracini is restricted here to the single included genus Neodiogmites.

7.5. Blepharepiini Papavero, 1973b.

Type genus Blepharepium Rondani, 1848. Genera included: Blepharepium. Autapomorphies: None. Apomorphic character states: Maxillary palpus one-segmented (11), ventral margin of postmentum of labium with median groove in distal part (19), prosternum and proepisternum fused and prosternum narrow above coxa (74; fig. 67), ventral metathoracic femora lacking macrosetae (116), metathoracic tibiae arched medially (119), crossvein r-m in distal half of discal cell (138), female spermathecae occupying more than posterior three abdominal segments (182), female furcal apodeme developed as short platelike apodeme (185), male S8 reduced, all other tergites and sternites well developed (193), male gonostyli positioned medially on gonocoxites (206), male ejaculatory apodeme triangular in anterior view (two lateral and one dorsal surface) (220). The Blepharepiini was described by Papavero (1973b) for the single Neotropical genus Blepharepium and the present analysis supports its morphological distinctness as no other genus groups with Blepharepium.

7.6. Molobratiini Lehr, 1999.

Type genus Molobratia Hull, 1958. Genera included: Molobratia. Autapomorphies: None. Additional apomorphic character states: Mystax extending over lower facial half (7), ventral margin of postmentum of labium entirely smooth (19), anterior ommatidia all same size (46), macrosetae on lateral margin of frons (dorsal to antennal insertion) present (49), apical “seta-like” sensory element situated apically on stylus (57), stylus composed of one element (58), postsutural dorsocentral setae absent (94), cell m₃ open (135), female T8 anteriorly lacking apodeme (169), female S8 platelike and slightly emarginate mediodistally (170), female without spurs on ovipositor (174), female spermathecal reservoir formed by more or less expanded and coiled ducts (180), male sperm sac appearing weakly sclerotized (218). Hull (1962) placed Molobratia in his Dasypogoninae: Dioctriini. The Molobratiini was described by Lehr (1999) originally in Dasypogoninae, but has been grouped in Dioctriinae based on the lack of acanthophorite spurs on the ovipositor and the shape of S8 by the same author (Lehr, 2001). The placement of Molobratia within Dioctriinae is not supported in the present study as it is positioned deeply within Dasypogoninae.

7.7. Megapodini Carrera, 1949.

Type genus Megapoda Macquart, 1834. Genera included: Cyrtophrys, Deromyia, Megapoda, Pegesimallus, Pseudorus, Senobasis. Autapomorphies: Two distinct macrosetae on median occipital sclerite (65; fig. 51). Additional apomorphic character states: Regular setae form ocellar setae (62), regular setae form antepronotum setae (69), one notopleural seta present (88), only regular setae on lateral metathoracic coxae (no macrosetae) (112), male epandrium and hypandrium at least partially fused (201). Megapodini was described by Carrera (1949), and Hull (1962) later labeled it as a separate subfamily taxon. This taxon has been extensively reviewed by Papavero (1975), who also provided a diagram of relationships of the included genera. Although his analysis is not phylogenetic, i.e., no distinction between apomorphic and plesiomorphic character states is made, a character matrix was provided and is re-analysed here (appendix 2). Figure 127 illustrates Papavero's hypothesis of relationships and figure 128 shows the result of the parsimony re-analysis, which results in three equally most parsimonious trees of 33 steps length (CI  =  0.78; RI  =  0.75) and a strict consensus topology of 35 steps (CI  =  0.74; RI  =  0.68). Megapoda, Pronomopsis Hermann, 1912, and Pseudorus form a clade in both diagrams, but the relationship between Cyrtophrys, Deromyia, Pegesimallus, and Senobasis is unresolved in the re-analysis. Lagodias Loew, 1858, has been synonymized with Pegesimallus by Londt (1980). Similar results to the re-analysis of Papavero's data are also found in the present analysis in that Cyrtophrys and Deromyia form a clade (Cyrtophryina), Megapoda and Pseudorus form a clade (Megapodina), and all four taxa, including Pegesimallus (Lagodiina) and Senobasis (Senobasina), are unresolved in a polytomy.

Figure 127–128

Phylogenetic relationships within Megapodini. 127. Original hypothesis proposed by Papavero (1975). 128. Phylogenetic re-analysis based on the characters of table 2 in Papavero (1975) summarized in appendix 2. The analysis in TNT results in three equally most parsimonious trees (length  =  33; CI  =  0.78; RI  =  0.75) and shown is strict consensus of these three trees (length  =  35; CI  =  0.74; RI  =  0.68). An outgroup was coded with state 0 for all characters to root the cladogram and is omitted here.

7.7.1. Megapodina Carrera, 1949.

Genera included: Megapoda, Pseudorus. Autapomorphies: None. Additional apomorphic character states: Pharyngeal pump with entire sclerotized ring (40), anatergal setae absent (82), only regular setae on posterior meron and metanepisternum (85), setae on superoposterior anepimeron present (86), metkatepisternum large and visible between mesothoracic and metathoracic coxae in ventral view (100), prothoracic and mesothoracic coxae distinctly oriented posteriorly (109), only one dorsal longitudinal ridge on pulvilli (124), cell r₁ closed, R₁ and R₂₊₃ as stalk vein present (131), R₄ terminating anterior to wing apex (143), male hypandrium well developed and rectangular or square in shape (199), male gonostyli positioned medially on gonocoxites (206), male dorsal aedeagal sheath long so that sperm sac is entirely enclosed (215).

7.7.2. Cyrtophryina Papavero, 1975.

Type genus Cyrtophrys Loew, 1851. Genera included: Cyrtophrys, Deromyia. Autapomorphies: None. Additional apomorphic character states: Lower facial margin slightly developed (4), pharyngeal pump with entire sclerotized ring (40), ventral metathoracic femora only setose, no macrosetae present (116), metathoracic tibiae arched medially (119), setiform empodium well developed (as long as claws) (127), alula reduced to small lobe (130), microtrichia on posterior wing margin arranged in two divergent planes (140). Papavero (1975) erected the taxon Cyrtophryina within his Megapodini for the two Neotropical genera, Cyrtophrys and Deromyia. The monophyly of this taxon is here corroborated.

7.7.3. Lagodiina Papavero, 1975.

Type genus Lagodias Loew, 1858  =  Pegesimallus Loew, 1858. Genera included: Pegesimallus. Autapomorphies: None. Apomorphic character states: Mystax extending over lower facial half (7), setae on dorsal postpedicel absent (55), apical “seta-like” sensory element situated apically in cavity on stylus (57), stylus composed of one element (58), postsutural dorsocentral setae absent (94), metathoracic trochanter with macrosetae and setae medially (114), cell m₃ open (135), R₄ terminating posterior to wing apex (143), female with setae on T8 directed anteriorly (168), male with marginal setae on T2 (186), male lacking gonocoxal apodeme (204), male with cerci fused to a single sclerite (212), male dorsal aedeagal sheath strongly sclerotized, appearing entirely closed (217). Pegesimallus is the type genus of Lagodiina and the only representative as no other Dasypogoninae genus groups with it.

7.7.4. Senobasina Papavero, 1975.

Type genus Senobasis Macquart, 1838. Genera included: Senobasis. Autapomorphies: None. Apomorphic character states: Lower facial margin slightly developed (4), two notopleural setae present (88), presutural dorsocentral setae present (92), only regular setae on metathoracic femora ventrally (no macrosetae) (116), metathoracic tibiae arched medially (119), setiform empodium well developed and as long as claws (127), dorsal surface of T2 smooth, without protuberances (154), female T10 composed of a single sclerite and entirely fused to T9 so that they are indistinguishable (172, 173), female spermathecal reservoir heavily sclerotized (181), male lateral ejaculatory process present and large cylindrical sclerite (213), male sperm sac appearing weakly sclerotized (218).

Unplaced genera: Archilestris, Diogmites, and Lestomyia. Archilestris, lacking a prothoracic tibial spine, is grouped within Stenopogoninae: Enigmomorphini by Artigas and Papavero (1991) and is here transferred to the Dasypogoninae, Diogmites is considered to be a Lastauracini (see above), and Lestomyia has been grouped within Dasypogonini by Papavero (1973b) (see above).

Phylogenetic relationships within Dasypogoninae: Lestomyia is the sister group to the remaining Dasypogoninae, which are characterized by the following apomorphic character states: 4: 4, 19: 2, 88: 3, 92: 0, 168: 1, 180: 1. The taxa Dasypogonini, Saropogonini, Thereutriini, Lastauracini and Diogmites, Blepharepiini, Molobratiini, and Megapodini form a grade of less inclusive sister taxa. The separate clades are delimited by a number of apomorphic character states, but no autapomorphies emerge from the present analysis.

8. Stenopogoninae

Stenopogoninae Hull, 1962. Type genus Stenopogon Loew, 1847. Species represented: 13. Figures 120–121, 124.

Autapomorphies: None. Additional apomorphic characters states: Facial swelling and mystax extending over lower facial half (4, 7), postmentum ventrally with median groove in distal part (19), all ommatidia same size (46), postpedicel cylindrical throughout (54), setae on anteroventral prothoracic tibiae absent (105), female spermathecal reservoir formed by more or less expanded and coiled ducts (180). Stenopogoninae sensu previous authors has been suspected to be non-monophyletic and used to be the largest taxon of Asilidae containing 118 genera (Geller-Grimm, 2004) of which 37 are included in the present analysis. It was divided into 10 subordinate taxa to accommodate this diversity in a meaningful way by previous authors. In the new delimitation of Stenopogoninae, only three of the subordinate taxa are recognized here and the remaining taxa are either placed as a higher-level taxon, i.e., Bathypogoninae, Phellinae, Tillobromatinae, and Willistonininae, or transferred to another higher-level taxon, i.e., Acronychini to Leptogastrinae, Ceraturgini and Cyrtopogonini to Brachyrhopalinae.

8.1. Enigmomorphini Hull, 1962.

Type genus Enigmomorphus Hermann, 1912. Genera included: Connomyia, Creolestes, Daspletis, Microstylum, Prolepsis. Autapomorphies: None. Additional apomorphic character states: Maxillary stipites fused medially, but V-shaped indentation present (14), maxillae proximally high and distal tip distinctly narrower (28), regular setae form discal scutellar setae (99), female spermathecae occupying at most three posterior segments (182). The Enigmomorphini was originally described by Hull (1962) to accommodate the Neotropical genus Enigmomorphus Hermann, 1912, and later expanded by Artigas and Papavero (1991) to include 14 genera. The New World genus Archilestris is transferred to the Dasypogoninae and the two Afrotropical genera Connomyia and Daspletis are transferred here to Enigmomorphini.

8.2. Plesiommatini Artigas and Papavero, 1991.

Type genus Plesiomma Macquart, 1838. Genera included: Plesiomma. Autapomorphies: None. Additional apomorphic character states: Vertex between compound eyes sharply depressed (3), facial swelling extending over entire face (4), mystax restricted to lower facial margin (7), dorsal flange medially on prementum absent (22), hypopharynx seta-like spicules spaced far apart (33; fig. 43), pharyngeal pump with entire sclerotized ring on lateral margin and circular in shape (40, 41), anterior tentorial pits well developed and conspicuous anteromedially located (44), frons at antennal insertion markedly approximating medially at level of antennal insertion (48), regular setae on lateral margin of frons dorsal to antennal insertion (no macrosetae) (49), regular setae forming postpronotum setae (no macrosetae) (70), scutellum small and mesopostnotum visible in dorsal view (97), apical scutellar setae absent (98), regular setae forming setation on ventral metathoracic femora (116), T2 longer than wide, but less than twice as long (153), female common spermathecal duct long, extending beyond tip of furca (177), male lateral ejaculatory process present and formed by a large cylindrical sclerite (213). The Plesiommatini was described by Artigas and Papavero (1991) to accommodate the three Neotropical genera Cystoprosopa (Hull, 1962), Dapsilochaetus (Hull, 1962), and Plesiomma.

8.3. Stenopogonini Hull, 1962.

Genera included: Gonioscelis, Ospriocerus, Scleropogon, Stenopogon. Autapomorphies: None. Additional apomorphic character states: Head circular, as wide as high, in anterior view (1; fig. 29), maxillary stipites divided medially (14), ventral margin of postmentum entirely smooth (19), dorsoposterior margin of cibarium with only one ridge connecting cornua and cornua not originating from ridge (35), cornua developed in anteroposterior orientation (37), prosternum and proepisternum separated and prosternum triangular with pointed dorsal tip (74), cell m₃ closed (135), R₂₊₃ apex arching sharply anteriorly in 90° angle (142), male hypopygium not rotated (195). The Stenopogonini is restricted here to only those robber flies with a circular head and with the character combination mentioned above.

Unplaced genera: The two widespread genera Ancylorhynchus and Scylaticus are part of the Stenopogoninae and form a clade supported by the following apomorphic character states, i.e., 62: 1, 74: 2, 213: 3, but do not group with any other described subordinate taxon.

Phylogenetic relationships within Stenopogoninae: Plesiommatini is the sister taxon to Stenopogonini and this clade is supported by the following apomorphic character states: setae on dorsal postpedicel present (55), presutural dorsocentral setae absent (92), sperm sac appearing more or less heavily sclerotized (218). The three clades, Enigmomorphini, Plesiommatini + Stenopogonini, and Ancylorhynchus + Scylaticus, form an unresolved polytomy and no comments on their relationship can be made.

9. Willistonininae, new status

Willistonininae Artigas and Papavero, 1991. Type genus Willistonina Back, 1908. Species represented: 5. Figures 120–121, 125.

Autapomorphies: None. Additional apomorphic character states: Only lower facial margin slightly developed (4), prosternum and proepisternum separated, prosternum triangular and pointed dorsally (74; fig. 68), discal scutellar setae absent (99), setae on legs dorsoventrally flattened (also circular setae present) (103). The taxon Willistoninini was described by Artigas and Papavero (1991) for the single Nearctic genus Willistonina. It is here expanded to include the Afrotropical genera Acnephalum and Trichoura, the Afrotropical and Palaearctic genus Sisyrnodytes, and the Nearctic genus Ablautus. Acnephalum and Sisyrnodytes form a clade characterized by a number of apomorphic character states (3: 2, 7: 3, 22: 1, 58: 1, 65: 1, 201: 0). The sister group to this taxon is composed of the remaining genera, which are characterized by the following apomorphic character states: 25: 0, 32: 0, 64: 1, 112: 2. Trichoura and Willistonina form a clade supported by the following character states: 7: 0, 8: 1, 48: 1, 49: 1, 76: 0, 113: 1, 130: 1. Because of its position as the sister group to a larger clade I, Willistonininae is recognized here as a higher-level taxon.

10. Dioctriinae

Dioctriinae Enderlein, 1936. Type genus Dioctria Meigen, 1803. Species represented: 9. Figures 120–121, 125.

Autapomorphies: None. Additional apomorphic character states: Maxillary stipites divided medially (14), ommatidia all same size (46), setae on dorsal postpedicel present (55), apical “seta-like” sensory element situated subapically laterally on stylus (57; fig. 55), cell cup open (136; fig. 83), male hypandrium and epandrium not approximating, separated by gonocoxites (201), subepandrial sclerite with lateral protuberances (211). The Dioctriinae, originally described by Enderlein (1936), was not applied as a higher-level taxon until re-erected by Lehr (2001). It is here shown to be a distinct taxon from other Dasypogoninae and Stenopogoninae sensu previous authors in which Dioctriini has been placed (e.g., Hull, 1962; Adisoemarto and Wood, 1975; Artigas and Papavero, 1991).

10.1. Dioctriini Enderlein, 1936.

Genera included: Dioctria, Eudioctria, Nannodioctria. Autapomorphies: Metathoracic coxae with pointed, peglike protuberance anteriorly (113). Additional apomorphic character states: Macrosetae on lateral margin of frons (dorsal to antennal insertion) absent (49), male epandrium composed of a single sclerite, fused more or less entirely medially (197). Dioctria and Nannodioctria form a less inclusive clade and it is characterized by two autapomorphies, i.e., cross section of prementum of labium square proximally (18), postmentum of labium platelike and arched (20; fig. 35), as well as additional apomorphic character states (23: 2, 24: 1).

10.2. Echthodopini Adisoemarto and Wood, 1975.

Type genus Echthodopa Loew, 1866. Genera included: Echthodopa. Autapomorphies: None. Additional apomorphic character states: Mystax extending over lower facial half (7), sensory pit in distal maxillary palpomere present (12), antepronotum with macrosetae and setae (69), setae on superoposterior anepimeron present (86), macrosetae present on ventral metathoracic femora (116), female S8 platelike and slightly emarginate mediodistally (170), female spermathecal reservoir formed by more or less expanded and coiled ducts (180), male subepandrial sclerite without lateral protuberances (211), male lacking lateral ejaculatory process (213), male dorsal aedeagal sheath long and sperm sac entirely enclosed (215), male ventral median margin of dorsal aedeagal sheath strongly sclerotized and appearing entirely closed (217).

Unplaced genera: Broticosia, Dicolonus, Myelaphus. The Australian genus Broticosia is found here to be the sister taxon to the remaining Dioctriinae. Dicolonus is characterized by an autapomorphy, i.e., female S7 reduced in size (166), and by a number of additional apomorphic character states (4: 3, 7: 2, 75: 1, 82: 1, 122: 0, 182: 1, 197: 0). The genus Myelaphus is nested deeply within Dioctriinae here as has been suggested before by Hull (1962) and Adisoemarto and Wood (1975). Artigas and Papavero (1991) placed it in Stenopogoninae: Ceraturgini along with Ceraturgus, which is here considered to be a Brachyrhopalinae.

Phylogenetic relationships within Dioctriinae: Broticosia is the sister group to the remaining Dioctriinae, which are characterized by these apomorphic character states: 19: 1, 49: 1, 53: 1, 88: 3, 118: 2, 130: 1, 193: 4, 218: 1. Dicolonus, Echthodopini, Myelaphus, and Dioctriini form a grade of less inclusive sister taxa. The separate clades are delimited by a number of apomorphic character states, but no autapomorphies emerged from the present analysis. A phylogenetic analysis of Dioctriini and Echthodopini based on Nearctic species, where these taxa are most diverse, was published by Adisoemarto and Wood (1975). The present analysis contradicts some of the hypotheses postulated by Adisoemarto and Wood in that Eudioctria is considered here to be a Dioctriini as it groups with Dioctria and Nannodioctria, which are considered to be the only representatives of Dioctriini by Adisoemarto and Wood. The Echthodopini are not recovered as monophyletic here because Dicolonus, Echthodopa, and Myelaphus do not form a clade. This taxon is therefore restricted to its type genus Echthodopa.

11. Leptogastrinae

Leptogastrinae Schiner, 1862. Type genus Leptogaster Meigen, 1803. Species represented: 7. Figures 120–121, 125.

Autapomorphies: Abdominal T2 more than twice as long as wide (153), S2 divided medially into two equal halves, which are separated by fenestra (159). Additional apomorphic character states: Seta-like spicules on hypopharynx spaced far apart (33; fig. 43), ocellar setae absent (62), only regular setae form katatergal setae (83), presutural dorsocentral setae absent (92), lateral depression on prothoracic coxae absent (104), row of macrosetae on anterodorsal mesothoracic tibiae absent (110), proximal prothoracic and mesothoracic tarsomeres longer than following two tarsomeres combined (120, 121), pulvilli absent (123), claws fairly straight throughout (125), R₄ and R₅ more or less parallel (133; fig. 85), R₄ terminating posterior to wing apex (143; fig. 85), female spermathecal reservoir formed by more or less expanded ducts to sac-shaped reservoir (180), female spermathecal reservoir heavily sclerotized (181), male lateral ejaculatory process present, large cylindrical sclerite (213; fig. 101). The monophyly of Leptogastrinae has never been in question and is corroborated here by many character states. Hull (1962) included the genus Acronyches within Leptogastrinae, but this placement was not acknowledged by Martin (1968). As Acronyches is the sister taxon to the Leptogastrinae sensu previous authors and two autapomorphies characterize this clade, I include this genus in the newly delimited Leptogastrinae. The relationships within this taxon will not be discussed here in detail as a comprehensive generic-level phylogeny is currently in preparation (Dikow, unpublished).

11.1. Acronychini Artigas and Papavero, 1991, new placement.

Type genus Acronyches Williston, 1908. Genera included: Acronyches. Autapomorphies: None. Additional apomorphic character states: Vertex sharply depressed (3), frons markedly approximating medially at level of antennal insertion (48), metkatepisternum large and visible between mesothoracic and metathoracic tibiae (100), setiform empodium minute or entirely absent (127), microtrichia on posterior margin of wing arranged in two divergent planes (141), female S8 platelike and hypogynial valves separated (surrounded by membrane) (170), female furca divided into two lateral sclerites (184).

11.2. Leptogastrini Schiner, 1862.

Genera included: Beameromyia, Euscelidia, Lasiocnemus, Leptogaster, Tipulogaster. Autapomorphies: Postpronotal lobes extended medially and anteriorly and nearly touching medially (72; fig. 65), metathoracic coxae directed anteriorly (111), male with surstylus on epandrium (198; fig. 101), male with lateral process of gonostyli present (207; fig. 101). Additional apomorphic character states: Only lower facial margin developed (4), dorsoposterior margin of cibarium with one transverse ridge connecting cornua (35), cornua on cibarium well developed in anteroposterior orientation (37), postpedicel tapering distally (54), postpronotum setae absent (70), anterior anepisternal and proepimeral setae erect to appressed and anteriorly directed (78, 79), postmetacoxal bridge partly present laterally and membranous area medially (102), cell cup open (136; fig. 85), female with short spermathecal ducts (178; fig. 99), male with gonocoxite or gonocoxite-hypandrial complex partially fused to epandrium (203; fig. 101), male lacking gonocoxal apodemes (204), male with sperm sac appearing more or less heavily sclerotized (218). The clade comprising Lasiocnemus, Leptogaster, and Tipulogaster is characterized by a single autapomorphy, i.e., only crossvein m-m closing cell d and M₁ branching of proximally (137; fig. 85), and additional apomorphic character states (177: 1, 199: 0, 202: 1).

12. Trigonomiminae

Trigonomiminae Enderlein, 1914. Type genus Trigonomima Enderlein, 1914. Species represented: 7. Figures 120–121, 125.

Autapomorphies: Anterior tentorial pits well developed and conspicuous anteroventrally located (44; fig. 26). Additional apomorphic character states: Apex of labella pointed (24), postocular setae formed by regular setae (no macrosetae) (66), prosternum and proepisternum fused, prosternum narrow above prothoracic coxae (74), supra-alar setae absent (90), female spermathecae occupying more than posterior three abdominal segments (182). The monophyly of Trigonomiminae has not been questioned previously due to the distinct appearance of these flies that are often referred to as “goggle-eyed” robber flies.

12.1. Trigonomimini Enderlein, 1914.

Genera included: Holcocephala, Rhipidocephala, Trigonomima. Autapomorphies: None. Additional apomorphic character states: Setae on dorsal postpedicel present (55), apical scutellar setae absent (98), T1 divided medially by membraneous area (152). Rhipidocephala is characterized by a single autapomorphy, i.e., female with S8 platelike and the distal margin is medially membraneous (170), and additional apomorphic character states (11: 1, 24: 2, 57: 3, 116: 1, 121: 1, 136: 0, 199: 1, 202: 1). Holcocephala is characterized by a single autapomorphy, i.e., face divided into a distinct ventral and dorsal half (5), and additional apomorphic character states (18: 2, 92: 0, 94: 0, 122: 0, 193: 4, 197: 2, 206: 1). Rhipidocephala is the sister group to the clade Holcocephala + Trigonomima, which is characterized by these apomorphic character states: 3: 2, 133: 0, 92: 0, 94: 0, 122: 0, 193: 4, 197: 2, 208: 1.

12.2. Xenomyzini Hardy, 1948.

Type genus Xenomyza Wiedemann, 1817  =  Damalis Fabricius, 1805. Genera included: Damalis. Autapomorphies: Postpedicel with proximal bulb and distal elongated part (54; fig. 56). Additional apomorphic character states: Antennal stylus reduced to apical “seta-like” sensory element, which is situated subapically in cavity on postpedicel (57, 58; fig. 56), postpronotal setae absent (73), metathoracic trochanter medially with posteriorly directed protuberance (115), short, stout, cone-shaped macrosetae ventrally on metathoracic femora (116), costa terminates at CuA₂ and A₁ (140), female with abdominal T10 lost or fused to cerci (172), male hypopygium rotated (195), male epandrium formed by a single sclerite (197), male lacking gonocoxal apodeme (204), as well as 18: 2, 114: 2, 137: 2, 138: 1. The two Afrotropical species Damalis annulata and D. monochaetes are the sister taxa to the Oriental Damalis sp. (Palatupana) and this Afrotropical clade is characterized by the lack of discal scutellar setae (99) and the short proximal prothoracic tarsomere, which is shorter than the two following tarsomeres combined (122).

Phylogenetic relationships within Trigonomiminae: Trigonomimini and Xenomyzini are adelphotaxa and all genera included in the present analysis are assigned to a subordinate taxon.

13. Stichopogoninae

Stichopogoninae Hardy, 1930. Type genus Stichopogon Loew, 1847. Species represented: 7. Figures 120–121, 126.

Autapomorphies: Posterior margin of compound eyes distinctly sinuate in ventral half (64; fig. 60). Additional apomorphic character states: Maxillary stipites divided medially (14; fig. 33), cornua on cibarium well developed in anteroposterior orientation (37), frons markedly and suddenly diverging laterally at level of antennal insertion (48; fig. 27), prosternum and proepisternum fused and prosternum broad above prothoracic coxae (74; fig. 66), anterior anepisternal setae absent (78), discal scutellar setae absent (99), proximal metathoracic tarsomere shorter than two following tarsomeres combined (122), female S8 platelike and hypogynial valves separated (surrounded by membrane) (170; fig. 97), male gonocoxites fused to hypandrium (202), male ventral margin of dorsal aedeagal sheath strongly sclerotized (appearing entirely closed) (217). The monophyly of Stichopogoninae has never been in doubt and has been recently corroborated by a phylogenetic analysis by Cannings (2002).

13.1. Stichopogonini Hardy, 1930.

Genera included: Lissoteles, Stichopogon. Autapomorphies: Male hypandrium and epandrium fused and connected by a narrow sclerite proximally (201; fig. 112), male subepandrial sclerite with short, stout macrosetae on the distal margin ventrally (209). Additional apomorphic character states: Mystacal setae tightly packed and forming a distinct shield (8), postpedicel tapering distally (54), one seta forming notopleural setae (88), medial setae on posterior scutum erect and directed dorsally to anteriorly (96), setae on anteroventral prothoracic tibiae absent (105), prothoracic and mesothoracic coxae directed distinctly posteriorly (109), blunt protuberance on anterior metathoracic coxae (113), male cerci fused to a single sclerite (212).

Phylogenetic relationships within Stichopogoninae: The genera Lasiopogon and Townsendia are not placed in any subordinate taxon. Lasiopogon is the sister group to the remaining Stichopogoninae and characterized by an autapomorphy, i.e., male subepandrial sclerite with short, stout macrosetae on its entire ventral surface (210), and additional apomorphic character states (7: 2, 46: 0, 55: 1, 98: 2, 143: 1, 181: 1, 195: 1). Townsendia is the sister taxon to Stichopogonini and the clade combining both taxa is characterized by a number of apomorphic character states (4: 1, 75: 1, 114: 2, 197: 0, 213: 0, 215: 1). Townsendia is excluded from Stichopogonini because it lacks the two autapomorphies of Lissoteles and Stichopogon and the taxon sampling within this group is very small. Cannings (2002) postulated a phylogeny of Stichopogoninae based on eight genera. The main findings are corroborated by the present analysis. In both studies Lasiopogon is found to be the sister group to the remaining Stichopogoninae and Lissoteles and Stichopogon are most closely related. In fact, Lissoteles renders Stichopogon paraphyletic in the present analysis, but because of the limited taxon sampling, this finding should not be taken as conclusive evidence for the non-monophyly of Stichopogon. Martin (1961), in revising Lissoteles, mentioned that both genera are morphologically very similar.

14. Brachyrhopalinae, new status

Brachyrhopalinae Hardy, 1926. Type genus Brachyrhopala Macquart, 1847. Species represented: 18. Figures 120–121, 126.

Autapomorphies: None. Additional apomorphic character states: Stylus composed of two elements (58), proximal prothoracic tarsomere longer than following two tarsomeres combined (120), alula reduced in size to small lobe (130), R₄ and R₅ more or less parallel (133). The species here summarized within Brachyrhopalinae are the remnant of the Dasypogoninae and Stenopogoninae sensu previous authors. It is a large clade that is not delimited by many unique apomorphic character states. Future phylogenetic studies will hopefully improve the delimitation of this taxon by including additional species that are morphologically similar to the Brachyrhopalinae representatives used here. From the valid and available family-group names represented in this clade, the name Brachyrhopalinae Hardy, 1926, with the type genus Brachyrhopala from Australia, is the oldest name and is chosen here to name this taxon.

14.1. Cyrtopogonini Artigas and Papavero, 1991, new placement.

Type genus Cyrtopogon Loew, 1847. Genera included: Afroholopogon, Cyrtopogon, Dasycyrton. Autapomorphies: None. Additional apomorphic character states: Female T8 with anterior rectangular apodeme, which is entirely fused to the tergite (169; fig. 87), female spermathecal reservoir formed by more or less expanded and coiled ducts (180). The Cyrtopogonini was described by Artigas and Papavero (1991) within their Stenopogoninae for 20 genera. Of these, only Cyrtopogon and Dasycyrton are recovered within this clade and Afroholopogon is referred to this taxon. Other genera originally included by Artigas and Papavero and also part of the present analysis are now found within Brachyrhopalinae, i.e., Heteropogon, Holopogon, Metapogon, Nannocyrtopogon, and Rhabdogaster, or transferred to other higher-level taxa, i.e., Ablautus to Willistonininae.

14.2. Chrysopogonini Hardy, 1934, new placement.

Type genus Chrysopogon von Röder, 1881. Genera included: Chrysopogon, Codula. Autapomorphies: Single, short, stout macroseta forms notopleural seta (88), cuticular facets on proximal metathoracic femora (117). Additional apomorphic character states: Prementum of labium laterally compressed (18), labellae of labium reduced and fused entirely to prementum (23), supra-alar, postalar, and presutural dorsocentral setae absent (90, 91, 92), alula well developed (130), abdominal tergites indented and setae with large sockets (151), female S8 proximally platelike and hypogynial valves forming a keel distally (170), female with T9–10 separated (not fused) (173), female without spurs on the ovipositor (174), female spermathecal reservoir formed by more or less expanded and coiled ducts (180), female furcal apodeme present and long (185). This exclusively Australian taxon was described by Hardy (1934) and was extensively reviewed by Clements (1985). Originally it included only the genera Chrysopogon, Codula, and Opseostlengis, but Hull (1962) later transferred Chryseutria to this taxon. The genus Codula was placed in the Stenopogoninae sensu previous authors in contrast to Chrysopogon, which was placed in Dasypogoninae sensu previous authors, in the most recent generic catalogue (Geller-Grimm, 2004) based on the absence of a prothoracic tibial spine. The close relationship between these two genera is supported here by autapomorphies and the placement of Codula within Chrysopogonini is without doubt.

14.3. Brachyrhopalini Hardy, 1926, new placement.

Type genus Brachyrhopala Macquart, 1847. Genera included: Austrosaropogon, Brachyrhopala, Cabasa. Autapomorphies: None. Additional apomorphic characters states: Lower facial margin slightly developed (4), discal scutellar setae absent (99). The Brachyrhopalini was described by Hardy (1926) for the only genus Brachyrhopala although he considered Cabasa to be a junior synonym of Brachyrhopala. It is expanded here to include Austrosaropogon.

14.4. Ceraturgini Artigas and Papavero, 1991, new placement.

Type genus Ceraturgus Wiedemann, 1821. Genera included: Ceraturgus, Nannocyrtopogon. Autapomorphies: None. Additional apomorphic character states: Facial swelling extending over lower facial three quarters (7), hypopharynx parallel-sided throughout and distal tip suddenly pointed (32), regular setae form katatergal setae (83), male hypopygium rotated (195), male cerci fused to single sclerite (212), male sperm sac appearing more or less heavily sclerotized (218). The Ceraturgini was described by Artigas and Papavero (1991) for the two genera Ceraturgus and Myelaphus. Myelaphus is transferred here to the Dioctriinae and Nannocyrtopogon, which is placed in Cyrtopogonini by Artigas and Papavero (1991), is placed in Ceraturgini here.

Additional genera of Brachyrhopalinae that do not group with any of the above-mentioned family-group taxa are: Cophura, Heteropogon, Holopogon, Leptarthrus, Metapogon, Nicocles, Rhabdogaster. Heteropogon, Holopogon, Metapogon, and Rhabdogaster were placed in Cyrtopogonini by previous authors, but do not group with Cyrtopogon in the present analysis. Cophura, Leptarthrus, and Nicocles were placed in the unavailable Isopogonini within Dasypogoninae by Papavero (1973b) based on the possession of an S-shaped spur on the prothoracic tibiae. These genera do not form a monophyletic clade in the present analysis and therefore a taxon cannot be recognized and I do not provide a new family-group name to replace the unavailable name. Although the genera Holopogon and Leptarthrus as well as Cophura, Heteropogon, Metapogon, Nicocles, and Rhabdogaster form clades, respectively, I do not propose any new family-group names for these taxa as I believe a more detailed study focusing on morphologically similar species will shed new light on the phylogenetic relationships within Brachyrhopalinae.

Phylogenetic relationships within Brachyrhopalinae: Cyrtopogonini is the sister group to the remaining Brachyrhopalinae and this clade is characterized by a number of apomorphic character states (22: 1, 32: 1, 83: 1). Holopogon and Leptarthrus form a clade delimited by these apomorphic character states: 7: 3, 90: 0, 168: 1, 181: 0, 201: 2. A clade comprising Brachyrhopalini, Ceraturgini, Chrysopogonini, and a number of unplaced genera is the sister group to Holopogon and Leptarthrus and characterized by two apomorphic character states (78: 0, 136: 0). Brachyrhopalini and Chrysopogonini form a clade delimited by a number of apomorphic character states (25: 1, 58: 3, 137: 2, 168: 0, 197: 0, 199: 0, 202: 1) that is restricted to Australia and its sister taxon is characterized by these apomorphic character states: 19: 1, 26: 0, 62: 2, 169: 1. Ceraturgini is the sister group to a clade comprising the five unplaced genera, which are united by a single apomorphic character state, i.e., female spermathecae occupying more than posterior three segments (182).

Incertae sedis

The two genera Coleomyia and Oligopogon could not be placed in any of the higher-level taxa and no family-group names are available for them. Coleomyia was placed in the Stenopogoninae: Tillobromatini sensu Artigas and Papavero (1991) and Geller-Grimm (2004), but it is not recovered within either taxon and is placed in an unresolved trichotomy with Dasypogoninae and Stenopogoninae. Oligopogon is placed as sister taxon to a large clade composed of Brachyrhopalinae, Dioctriinae, Leptogastrinae, Stichopogoninae, and Trigonomiminae. It had been placed in Stenopogoninae and Trigonomiminae before (see Geller-Grimm and Hradský, 2003, for a summary).

The following monogeneric family-group taxa Archilaphriinae (Archilaphria), Dasylechiini (Dasylechia), and Neophoneini (Neophoneus) are not represented in the present analysis and no comment on their phylogenetic relationships can be made. Archilaphria was placed within Dasypogoninae sensu previous authors and Dasylechia and Neophoneus in Laphriinae (Papavero, 1973a; Geller-Grimm, 2004).

Phylogenetic Relationships of Higher-Level Taxa within Asilidae

Figure 120

The most outstanding discovery of the present extensive morphological phylogenetic study is the hypothesis that Laphriinae is the sister taxon to the remaining Asilidae. This relationship has never before been suggested and Laphriinae is generally assumed to be phylogenetically derived Asilidae. Morphologically, however, Laphriinae shares a number of symplesiomorphies with Asiloidea outgroup taxa. In most Asilidae, the prothoracic and mesothoracic coxae are oriented ventrally, which probably facilitates the specialized capture and grasping position of prey items directly underneath the thorax. In Laphriinae except Atomosiini and outgroup taxa, the pro- and mesothoracic coxae are oriented posteriorly, which is the plesiomorphic character state within Asilidae. This state is also found in a few robber-fly species phylogenetically unrelated to Laphriinae. Another interesting character in Laphriinae pertains to the male terminalia. Yeates (1994) and Sinclair et al. (1994) independently proposed the proximal connection of the epandrium and hypandrium as a possible autapomorphy of Asilidae because it is found in most assassin flies. Both studies considered only a small representation of Asilidae, and species of Laphriinae were not represented. In Laphriinae, the hypandrium is generally very small, or even absent, and therefore not connected proximally to the epandrium (figs. 102, 108–109). This condition is homologous to the state found in the included Asiloidea outgroup taxa. Therefore, this character state is autapomorphic for a less inclusive taxon within Asilidae indicated as taxon A in figures 120–121. This clade is characterized by the following apomorphic character states: prosternum and proepisternum separated and the prosternum is square to rectangular and straight dorsally (character 74; fig. 69), prothoracic and mesothoracic coxae more or less directed ventrally (sometimes even anteriorly) (109; fig. 61), female furcal apodeme present, long and platelike in shape (185), hypandrium and epandrium approximating, but not fused proximally (201; fig. 113).

The Asilinae and Ommatiinae form a sister-group relationship (clade B) based on the following apomorphic character states: maxillary stipites fused medially, but V-shaped indentation present (14), tip of maxillae blunt (29; fig. 31), setae on posterior basalare present (81), setae on superoposterior anepimeron present (86), metkatepisternal setae present (101), female spermathecal reservoir not differentiated from spermathecal ducts (180), female spermathecal reservoir weakly sclerotized (181), male hypandrium well developed and rectangular or square (199; figs. 106, 113), male lateral ejaculatory process absent (213; fig. 114). Previous authors have assumed the two taxa were closely related due to their overall morphological similarity (Karl, 1959; Hull, 1962; Papavero, 1973a; Lehr, 1969, 1996).

Taxon C, i.e., Asilidae with the exception of Asilinae, Laphriinae, and Ommatiinae, is characterized by these apomorphic character states: vertex only slightly depressed (3; fig. 27), facial swelling and mystax extending over lower three quarters (4, 7), postpedicel medially broadest and long (54; fig. 52), setae on female T8 with erect setae (168), female S8 platelike with hypogynial valves extending (170), female T10 with acanthophorite spurs (174; figs. 87, 91), male sperm sac entirely free and sheath short (215; fig. 115), male aedeagal sheath weakly sclerotized ventrally and appearing incompletely closed (217; fig. 115). The next inclusive taxon D is characterized by these apomorphic character states: hypopharynx parallel-sided in proximal half and distal half much narrower (32; fig. 41), pharyngeal pump without sclerotized ring laterally (40), metathoracic coxae with regular setae laterally (no macrosetae) (112), female T9 and T10 only partly fused (173). Taxon E is characterized by the following apomorphic character states: sensory pit in distal maxillary palpomere present (12), anterior ommatidia all same size (46), cell m₃ open (135), female furcal apodeme short and platelike (185), male hypandrium and gonocoxites entirely free (202).

Dasypogoninae, Stenopogoninae, Tillobromatinae, and the unplaced genus Coleomyia form a clade (taxon F) characterized by the following apomorphic character states: dorsal flange on prementum of labium present (22), apical “seta-like” sensory element situated apically in cavity on stylus (57), cell cup open (136), male hypopygium rotated (195), male cerci fused to a single sclerite (212). Tillobromatinae is the sister group to the remaining two taxa plus Coleomyia (taxon G), which is characterized by the presence of macrosetae and setae on the postpronotum (70), postpronotal lobes with macrosetae and setae (73), anterior anepisternal setae absent (78), discal scutellar setae absent (99), lateral metathoracic coxae with macrosetae and setae (112). Dasypogoninae, Stenopogoninae and Coleomyia are part of an unresolved polytomy and the phylogenetic relationships and placement of Coleomyia need to be addressed with a denser taxon and character sampling informative for this particular group.

Clade H is characterized by prothoracic and mesothoracic tarsomeres shorter than following two tarsomeres combined (120, 121), discal cell closed by M₂ and m-m in a single line (137). Taxon I includes the unplaced genus Oligopogon, and clade J and is characterized by the following apomorphic character states: pharyngeal pump laterally with sclerotized ring, but medio-anteriorly unsclerotized (40; fig. 49), metathoracic trochanter medially with regular setae (no macrosetae) (114), setiform empodium reduced in length to varying degree (127), alula reduced entirely so that the wing margin is nearly straight (130; fig. 85). Oligopogon is the sister taxon to a larger clade J, which is delimited by: mystax restricted to lower facial margin (7), hypopharynx parallel-sided throughout and only distal tip suddenly pointed (32; figs. 39–40), ocellar setae formed by regular setae (no macrosetae) (62), antepronotum with regular setae (no macrosetae) (69), apical scutellar setae formed by regular setae (no macrosetae) (98). This clade includes four taxa previously thought to be monophyletic as well as the Brachyrhopalinae, which is a clade in which several genera previously assigned to Dasypogoninae and Stenopogoninae sensu previous authors are united.

Taxon K, which comprises Dioctriinae, Leptogastrinae, and Trigonomiminae, has never been postulated before, but is here supported by the following apomorphic character states: facial swelling extending over lower facial half (4), female T8 with posteriorly directed setae (168), female T9 and T10 entirely fused, not discernible (173), female T10 without acanthophorite spurs (174). Leptogastrinae and Trigonomiminae together form the sister group (taxon L) of Dioctriinae and are delimited by these apomorphic character states: maxillary palpus one-segmented (11), female S8 platelike and slightly emarginate mediodistally (170), male hypandrium reduced or not differentiated (199), male gonocoxite and hypandrium fused to form a gonocoxite-hypandrial complex (202).

The sister group to taxon K is a clade comprising Brachyrhopalinae and Stichopogoninae (taxon M). This taxon is only supported by a single apomorphic character state, i.e., maxillae (in lateral view) proximally high and distal tip distinctly narrow (28; fig. 38), and future studies with an increased taxon sampling particularly within the Brachyrhopalinae will hopefully provide more substantial morphological support.

Previous versus the Present Phylogenetic Hypothesis for Asilidae

As outlined in the introduction, Hull's (1962) and Papavero's (1973a) classifications and postulated diagrams of relationships are not phylogenetic in nature and cannot be compared to the present cladistic hypothesis. Karl's (1959; fig. 129) study is phylogenetic, but the limited taxon sampling especially in his “übrige Dasypogoninae,” which entails a significant number of species, makes a meaningful comparison difficult. Karl does not include Atomosiini in Laphriinae nor Lasiopogon in Stichopogoninae in contrast to the findings of the present study. The Leptogastrinae and Trigonomiminae (his “Damalis Gruppe”) are unresolved within Asilidae and do not form a monophyletic clade as postulated here. Karl pointed out, though, that Asilinae and Ommatiinae are sister taxa and Laphriini and Hoplistomerini are closely related, findings that have been corroborated here. The re-analysis of the character matrix provided by Karl (1959: fig. 47) results in a single most parsimonious tree (length  =  26; CI  =  0.69; RI  =  0.78; fig. 129, appendix 3). The topology is identical to the one illustrated by Karl in his figure 49, however, those relationships labeled as questionable by Karl are resolved. In summary, a clade Asilinae and Ommatiinae is the sister taxon to Leptogastrinae, Damalis is the sister group to Laphriinae (including Atomosia, Hoplistomerini, and Laphriini), and Lasiopogon and Stichopogon do not form a monophyletic clade.

Figure 129

Phylogenetic re-analysis of relationships of Asilidae based on the characters of figure 47 in Karl (1959) and summarized in appendix 3. The analysis in TNT results in a single most parsimonious tree (length  =  26; CI  =  0.69; RI  =  0.78). An outgroup was coded with state 0 for all characters to root the cladogram and is omitted here.

The only phylogenetic analysis of Asilidae representing most of the higher-level taxa sensu previous authors is the molecular study published by Bybee et al. (2004; fig. 130). The results are limited in their applicability because the analysis is based only on 26 Asilidae species and Ommatiinae, Stichopogoninae, Trigonomiminae, as well as Laphystiinae sensu previous authors are represented by only a single species, so that their monophyly could not be tested. However, the results can be compared to the present morphological study. Bybee et al. (2004) published two cladograms, one based on parsimony as the optimality criterion and the other based on maximum likelihood, without stating which is the preferred hypothesis. Although both hypotheses depict somewhat similar relationships there are differences in the position of certain species. I will base my discussion solely on the parsimony results (Bybee et al., 2004: fig. 3), which is illustrated in fig. 130, as the present morphological analysis is also based on this optimality criterion.

Figure 130

Previously published phylogenetic hypothesis of Asilidae by Bybee et al. (2004: fig. 3) based on parsimony analysis under direct optimization in POY of 16S, 18S, and 28S rDNA and COII DNA sequences.

The most striking difference between the phylogenetic hypothesis postulated here and the hypothesis of Bybee et al. (2004) is the position of the Leptogastrinae (figs. 120, 130). When Martin (1968) elevated the Leptogastrinae to family status, he regarded them as the sister group to Asilidae. This hypothesis is followed by Papavero (1973a), who excluded these slender flies from his exhaustive review of the Asilidae of the world. Oldroyd (1969), on the other hand, refuted Martin's study and emphasized that Leptogastrinae should be retained in Asilidae, although his arguments are not based on phylogenetic considerations. In the molecular study by Bybee et al. (2004), the Leptogastrinae groups as the sister taxon to the remaining Asilidae, which would support Martin's view and potentially allow for the separation of Leptogastrinae as a separate family taxon. However, in the present extensive morphological phylogeny, the Leptogastrinae is positioned deeply within Asilidae as sister taxon to the Trigonomiminae. I have much more confidence in the present grouping as a suite of apomorphic character states supports this hypothesis (fig. 120). Additionally, my own combined analysis of the morphological matrix presented here and DNA sequence data of five nuclear genes also supports a grouping of Leptogastrinae deeply within Asilidae (Dikow, submitted).

The higher-level relationships within Asilidae postulated by Bybee et al. (2004) are in general very different from the present hypothesis as is the monophyly or non-monophyly of certain higher-level taxa. Laphriinae is supported as monophyletic in the molecular study, but the genus Laphystia is not placed close to this group, and is considered the sister group to Stenopogonini. In the present study the Laphriinae sensu previous authors is also supported as monophyletic, but Laphystia is closely related to it and is part of the revised taxon Laphriinae. Apocleinae and Asilinae sensu previous authors are found to be paraphyletic with respect to each other in the present analysis, but are supported as monophyletic by Bybee et al. Furthermore, Ommatiinae is more closely related to Asilinae sensu previous authors in the molecular analysis, in contrast to it being the sister group to the revised Asilinae including the Apocleinae sensu previous authors as supported here. The Dasypogoninae is considered non-monophyletic in the molecular study, but the three genera included by Bybee et al., Diogmites, Saropogon, and Senobasis, are part of the Dasypogoninae sensu stricto in the present morphological study. The Stenopogoninae sensu previous authors is non-monophyletic in Bybee et al. and representatives appear in four different clades. The grouping Creolestes and Obelophorus as a clade is at odds with the findings of the present study, in which these two genera are placed in separate and phylogenetically unrelated higher-level taxa, i.e., Creolestes in Stenopogoninae: Enigmomorphini and Obelophorus in Phellinae. One must conclude that a DNA sequence study based on limited exemplars of a hyperdiverse insect taxon is a misguided exercise.

Biology of Asilidae

The newly proposed hypothesis that the Laphriinae constitutes the sister group to the remaining Asilidae makes a discussion of some biological phenomena possible. Larvae of Asilidae were described very early (see summary in Melin, 1923), but to date we know larvae of only a very few species and have limited knowledge of larval morphology and habitats. The most commonly encountered larvae of robber flies are those of Laphria and relatives, which are found within decaying wood in forest habitats where the majority of Laphriinae sensu previous authors occur. Based on this knowledge, one could assume that all larvae of Laphriinae develop in wood in contrast to all other known robber-fly larvae, which have been observed to develop in soil and sandy habitats (e.g., Melin, 1923; Fisher, in press). The new delimitation of Laphriinae includes all of those species previously assigned to Laphystiinae of which a single species is known in the larval stage. The larva of Laphystia carnea Hermann, 1906, was described by Krivosheina (1973) from Turkmenistan and, in contrast to other Laphriinae, the larva develops in sandy loam. Among the outgroup taxa, some Mydidae larvae are known to occur in wood whereas others, together with all Apioceridae larvae described so far, develop in sandy soil (e.g., Wilcox, 1981; Peterson, 1981). With this limited knowledge about larvae, one cannot, at this stage, postulate whether soil or wood is the ancestral larval habitat for Asilidae.

A similar scenario can be envisioned for the evolution of mimicry within Asilidae. Many Laphriinae species closely resemble various Aculeata bees and wasps (Hymenoptera). The earliest studies on mimicry in Asilidae were published by Brauer (1885) and Poulton (1904), who summarized some of the early observations by naturalists, and later Melin (1923) considered the problem in more detail. As Batesian mimics, robber flies can avoid predators such as birds. Furthermore, assassin flies that resemble their prey species can approach their prey items without being detected as was observed by Poulton (1904) or remain unnoticed at their perching site (Grimaldi and Engel, 2005). Many Laphriinae have been shown to feed on Aculeata although not exclusively (e.g., Londt, 1995, 2006b) as the majority of assassin flies are not specialized hunters. Mimicry of Hymenoptera is not restricted to Laphriinae, though, as some genera included in the present analysis, e.g., Megaphorus (Asilinae), Blepharepium and Diogmites (Dasypogoninae), and Ospriocerus (Stenopogoninae), have been observed to mimic Aculeata in Arizona and New Mexico (Linsley, 1960). Poulton (1904) observed Dasypogon diadema (Dasypogoninae) mimicking Anoplius viaticus Linnaeus, 1758 (as Pompilus; Pompilidae), in Europe. Enderlein (1914) described three genera, i.e., Ammophilomima, Ophionomima (both Leptogastrinae), and Trigonomima (Trigonomiminae), and named them based on the hypothesis that they mimic species of Ammophila Kirby, 1798 (Sphecidae), Ophion Fabricius, 1798 (Ichneumonidae), and Trigona Jurine, 1807 (Apidae), respectively. Londt (1991) described a bee-mimicking genus from southern Namibia in yet another clade, i.e., Stenopogoninae sensu previous authors. Some Mydidae, e.g., Mydas clavatus and Opomydas townsendi included in the present study, resemble Pompilidae and Vespidae wasps, respectively. Hesse (1969) described Vespiodes from southern Africa and highlighted the observation that species of this genus together with Mydaselpis Bezzi, 1924 (Syllegomydinae), resemble Vespidae wasps in this biodiverse region of Africa. In the arid southwestern United States and adjacent Mexico a robber fly, Wyliea mydas (Brauer, 1885) (Asilinae), and similar species of Mydas Fabricius, 1794, both mimic sympatrically occurring Pepsis Fabricius, 1804, wasps (Pompilidae) with an entirely black body and bright orange-red wings (Brauer, 1885). In contrast, Apiocera species are not known to mimic any other distasteful or dangerous insects. Based on the phylogenetic relationships of the above-mentioned taxa, one can postulate with certainty that mimicry has evolved independently several times within Asilidae. Only a more thorough study of Batesian mimicry of Asilidae and Mydidae species and a more comprehensive phylogenetic study including many more assassin-fly species can shed light on the question.

Biogeographical Implications

Although many of the higher-level taxa proposed previously and corroborated here are worldwide in distribution, some of the clades are biogeographically restricted (fig. 131Figure 131). Laphriinae, the sister taxon to the remaining Asilidae, is worldwide in distribution as are Asilinae, Dasypogoninae, Leptogastrinae, Ommatiinae, Trigonomiminae, Stenopogoninae, and Stichopogoninae (figs. 120, 131). The extant representatives of Bathypogoninae, the sister group to a large clade D, are restricted to Australia (and possibly South America) and representatives of Phellinae, sister group to clade E, occur only in Australia and Chile (figs. 120, 131). Extant Tillobromatinae are found in southern Africa and South America whereas Willistonininae is distributed in the Afrotropical and Nearctic regions and Sisyrnodytes also occurs in the Palaearctic Region (figs. 120, 131). Dioctriinae are most speciose in the Northern Hemisphere, but Broticosia, the sister taxon to the remaining Dioctriinae, is found only in Australia (figs. 121, 131). Brachyrhopalinae is found throughout the world, but the Brachyrhopalini and Chrysopogonini are restricted to Australia (figs. 121, 131). Although I attempted to include representatives from all zoogeographical regions the Oriental Region is certainly underrepresented. This is partly due to the lack of well-organized collections from which specimens can be borrowed, but also due to the lack of expertise and interest in this particular fauna. Perhaps most interesting is the repeated restriction of adelphotaxa to hyperdiverse radiations to areas in the Southern Hemisphere, especially Australia and southern South America, i.e., Bathypogoninae, Phellinae, and Tillobromatinae (figs. 120, 131).

Figure 131

Strict consensus topology as in figure 118, but taxon names substituted with zoogeographical region and country of origin of specimens examined. Arrow indicates root of Asilidae.

Figure 131

Continued.

Some Morphological Characters Revisited

Not all of the autapomorphies previously postulated for Asilidae by Hennig (1973) and Woodley (1989) could be corroborated in the present analysis. Hennig (1973) attributed a proximally constricted abdomen to Hull (1962), but I am unable to find any reference to this feature in Hull's publication. This character does not hold up as autapomorphic because in most Asilidae T2–3 are more or less square or rectangular, and only in a few species of Brachyrhopalinae, Dasypogoninae, Dioctriinae, Laphriinae, and Ommatiinae is the abdomen constricted, or waistlike (see character 156). Hennig (1973) also postulated densely bristled raptorial legs as autapomorphic for Asilidae, and although many species possess macrosetae on the legs, they could not be coded as such because the setal positions vary greatly among the species, making homologizing them difficult. Only a row of macrosetae on the anterodorsal mesothoracic tibia is coded here (see character 110) because it is found in most Asilidae. However, these setae are also present in most outgroup taxa and therefore not autapomorphic for assassin flies.

Oldroyd (1974a) and Woodley (1989) postulated the presence of the mystax as an autapomorphy, and in the present analysis it is absent in only Pseudorus distendens (Dasypogoninae) among Asilidae. However, macrosetae are also present on the face of Mydidae, as mentioned by Hardy (1927a) and Woodley (1989) although neither author homologized these setae with the ones in robber flies. I code them to be homologous as they are found at the same position on the head and cannot be separated otherwise (see character 6). The character optimization for the mystax is ambiguous at the root of the Asilidae and there are three equally parsimonious explanations for the character transformation. As the mystacal setae are absent in Apioceridae, Neorhaphiomidas (Mydidae), and Pseudorus, they could have originated (1) in the most recent common ancestor of Asilidae + (Apioceridae + Mydidae) and then become lost independently in Apioceridae, Neorhaphiomidas, and Pseudorus (4 steps, fast optimization), (2) independently in Asilidae and Mydidae and are reduced once in each of these taxa (4 steps, slow optimization), or (3) independently in Rhaphiomidas, in Mydidae except Neorhaphiomidas and Rhaphiomidas, and Asilidae and are reduced in Pseudorus (4 steps, unambiguous optimization). One morphological difference is evident in that Mydidae, with the exception of Mydas clavatus and Rhaphiomidas maehleri, possess an area medially on the face from which the mystacal setae are absent. I am unable to comment on this feature and a more detailed analysis of mystacal setae within Mydidae might shed light on the homology to the mystacal setae in Asilidae.

Hull (1962) and Papavero (1973a) discussed several morphological characters that were believed to contain important phylogenetic information for classifying Asilidae. Some of these will be discussed briefly in light of the phylogenetic hypothesis presented here, with the character transformation from the plesiomorphic to the apomorphic condition outlined.

Both authors argued that the open wing cells r₁, r₅, m₃, and cup represent the plesiomorphic condition within Asilidae. That this is not the case is exemplified, for example, by clade L (Leptogastrinae + Trigonomiminae) (fig. 120), which is a phylogenetically derived clade and in which all wing cells are generally open (only cell cup is closed in Acronyches and Lasiocnemus within Leptogastrinae and all Trigonomiminae except Rhipidocephala) (figs. 84–85). Furthermore, cells r₁, m₃, and cup are closed in Asilinae, Laphriinae (r₁ is open in Psilocurus), and Ommatiinae (figs. 73–80) as well as Apioceridae and Mydidae as the sister group, but the same cells are generally open in taxon C, i.e., the remaining Asilidae (figs. 81–83, 120). Cell m₃ and cup are also closed in Bathypogoninae and Phellinae, which are representatives of clade C, and open in the majority of species of taxon E (fig. 120). This indicates that the closed condition of these wing cells is actually the plesiomorphic state. The only exception is cell r₅, which is closed in very few Asilidae (figs. 73, 75) and Asiloidea outgroup taxa, and the plesiomorphic condition is an open cell r₅.

Previous classifications emphasized the one- or two-segmented maxillary palpi as an important character (e.g., Hull, 1962; Papavero, 1973a). Hull (1962) used this character in the first couplet to his identification key to higher-level taxa and Oldroyd (1974a: 11) stated, “It seems to me, therefore, that the number of palpal segments is too elusive and enigmatic a character to be given such prominence.” Hennig (1973) and Woodley (1989) postulated a two-segmented maxillary palpus as an autapomorphy for the Brachycera. It has been assumed that a two-segmented palpus is the plesiomorphic condition within Asilidae (Hull, 1962; Papavero, 1973a; Oldroyd, 1974a) and this hypothesis is corroborated here as this state is optimized at the root of Asilidae. It has been independently reduced to a single palpomere at least 13 times in all Asilinae, Bathypogoninae, Leptogastrinae, and Ommatiinae and at least some taxa of Dasypogoninae, Dioctriinae, Laphriinae, Stenopogoninae, Stichopogoninae, Trigonomiminae, and Willistonininae (see character 11).

Another character Hull (1962) and Papavero (1973a) emphasized is the development of the prosternum and its lateral fusion to the proepisternum. Clements (1951: 14) studied many species in detail and concluded that, “Examination of the prosternum in all groups showed that too much variation occurs for it to be used as a diagnostic or supplementary character.” There is no doubt that a diversity of shapes of the prosternum exists among assassin flies, but they are of value and can be easily delimited in five distinct character states (see character 74; figs. 66–69). Hardy (1948), Hull (1962), and Papavero (1973a) concluded that a complete broad prosternum, which is entirely fused to the proepisternum, is the plesiomorphic condition within Asilidae. This hypothesis is corroborated here as this character state is optimized at the root of Asilidae and a separated prosternum is an apomorphic state for taxon A. A complete prosternum is found in all Mydidae and those Laphriinae which form the basalmost clades, i.e., Hoplistomerini, Laphystiini, Perasis, Protometer, Psilocurus, and Zabrops, as well as in all Stichopogoninae and most Leptogastrinae in a phylogenetically derived position (fig. 66).

Adisoemarto and Wood (1975) considered the acanthophorite spurs (character 174; figs. 87, 91, 97) on the female ovipositor to be a ground-pattern feature of Asiloidea and Griffiths (1994) recognized them as an autapomorphy of this taxon. Yeates (1994) postulated that these spurs have actually been derived independently within Bombyliidae and the remaining Asiloidea. Papavero (1973a: 231) stated with regard to Asilidae that

The presence of acanthophorite spurs is an apomorphic character state for taxon C because the spurs are absent in Asilinae, Laphriinae, and Ommatiinae whereas they are present in a less inclusive clade within Asilidae (fig. 120). There are two equally parsimonious explanations for the character transformation. Because Apioceridae + Mydidae, as the sister group to Asilidae, possess acanthophorite spurs, these spurs (1) are reduced in the most recent common ancestor of Asilidae and then originated as a novel structure in the most recent common ancestor of taxon C (2 steps, fast optimization) or (2) are independently reduced in Laphriinae and taxon B (Asilinae + Ommatiinae), respectively (2 steps, slow optimization). Acanthophorite spurs are also absent in taxon K (Dioctriinae + (Leptogastrinae + Trigonomiminae)) and the absence is an apomorphic character state for this speciose clade. This highlights the fact that it is important not to consider the absence of spurs as the result of a single reduction, but to allow for independent losses among phylogenetically unrelated taxa of Asilidae. In contrast to Papavero's opinion (1973a), this character is therefore not of great importance to decipher the early cladogenesis of Asilidae. Based on the taxon sampling within the outgroup taxa, it appears that the spurs have also been independently derived in Bombylius major (Bombyliidae), Prorates sp. (El Hacha) (Scenopinidae), and Phycus frommeri (Therevidae).

Papavero (1973a: 231) postulated in reference to the presence of projections on the prothoracic tibiae that, “This character is very probably derived and monophyletic [sic], as all the steps in its transformation can still be traced. Genera presenting this character were segregated under one subfamily.” The pattern of evolution of this feature occurred differently based on the present phylogenetic hypothesis. The large spine on the prothoracic tibiae is found in Dasypogoninae as well as the Australian Brachyrhopalini and Chrysopogonini (Brachyrhopalinae) (figs. 70, 121). A small S-shaped spur, believed to be the intermediary stage by Papavero (1973a; fig. 72), is present in Thereutria in the Dasypogoninae and Cophura, Leptarthrus, and Nicocles of the Brachyrhopalinae. This highlights the fact that the S-shaped spur of Thereutria is homologous to the large spine of other Dasypogoninae although the shape has been transformed, but not homologous to the S-shaped spur found in Cophura, Leptarthrus, and Nicocles. In fact, the present hypothesis based on 220 characters rejects the homology of the S-shaped spur even among these three genera as they do not form a clade. Furthermore, the large spine of Dasypogoninae is not homologous to the one found in the Australian taxa despite its morphological similarity. Interestingly, Archilestris groups within Dasypogoninae based on a suite of morphological characters, but was excluded from this taxon by Artigas and Papavero (1991) because of the absence of tibial projections, which is postulated here to be the result of the reduction of this feature. It is shown here that a taxon based on only the presence of tibial projections is not recognizable as monophyletic, and the delimitation of Dasypogoninae is therefore adjusted.

Martin (1968) examined the lack of pulvilli in Diptera in detail to support the hypothesis that his Leptogastridae is more plesiomorphic than the remaining Asilidae and should therefore be grouped as a separate family taxon. He stated, “On the basis of the above discussion I suggest that the dipterous tarsal segment without pulvilli is primitive, and not homoplastic as Janssens hypothesized when he used the term convergent” (Martin, 1968: 81). This hypothesis is not corroborated here because the pulvilli are present in all outgroup taxa as well as most Asilidae, and are reduced only two independent times within Asilidae, i.e., once in the most recent common ancestor of Leptogastrinae and probably once in the most recent common ancestor of a subordinate taxon of Willistonininae, as not all Willistonininae lack pulvilli. Oldroyd (1969) also provided a detailed rebuttal of the lack of pulvilli in Diptera in response to Martin's study.

Transformation Series Within Multistate Characters

Thirty-one of the 106 multistate characters could be considered to be part of an evolutionary transformation series in contrast to the remaining multistate characters, which are coded in this way for convenience only as they deal with a particular character system. One should keep in mind that any multistate character can be recoded into a suite of binary characters without losing any information. Although these 31 “true” multistate characters could have been made additive to restrict the transformation from one state to another, all characters are treated as nonadditive so that no constraints on evolutionary transformation sequences are imposed. The “true” multistate characters are: 3 (vertex shape), 14 (stipites fusion), 19 (postmentum development), 23 (labella), 33 (hypopharynx dorsal protuberances), 40 (pharyngeal pump), 58 (stylus), 62 (ocellar setae), 64 (compound eye margin), 65 (macrosetae on median occipital sclerite), 66 (postocular setae), 69 (antepronotal setae), 70 (lateral postpronotal setae), 72 (postpronotal lobes), 73 (dorsal postpronotal setae), 83 (katatergal setae), 98 (apical scutellar setae), 102 (postmetacoxal bridge), 110 (setae on mesothoracic tibia), 112 (setae on metathoracic coxa), 114 (setae on metathoracic trochanter), 118 (fringe of parallel setae on metathoracic femur and tibia), 127 (empodium), 130 (alula), 153 (T2 length), 173 (T9–10 fusion), 175 (spermathecae number), 181 (spermathecal reservoir), 202 (hypandrium and gonocoxite fusion), 206 (gonostyli position), and 214 (aedeagus prongs). Table 8 summarizes those 20 multistate characters from the above list in which a transformation, for example, from state 0 to state 2 or vice versa occurs while the intermediate state 1 is not observed (based on the character optimization on the most parsimonious topology in fig. 118Article Foldout). All of the remaining 11 characters exhibit an additive transformation including all character states.

Table 8

Summary of transformation of character states in 20 multistate characters based on most parsimonious topology in figure 118 All optimizations unambiguous except where indicated with slow and fast

When the 20 multistate characters of table 8 are recoded to be additive the tree length of the most parsimonious topology increases from 2760 to 2834 steps. Many of the characters are associated with setal development where either no setae (state 0), regular setae (state 1), or macrosetae (state 2) are developed. When coded nonadditive, i.e., counting only one step from state 0 to state 2, it is shown that macrosetae can evolve without the prior existence of regular setae. Similarly, a complete postmetacoxal bridge (character 102) has evolved three independent times within Asilidae from an entirely membraneous postmetacoxal area (transformation from state 0 to state 3). These three phylogenetically independent occurrences of this particular transformation series indicate that the same pattern can be derived by different evolutionary processes if one acknowledges that no fixed genetic basis is responsible for the development of a postmetacoxal bridge.

The development of the posterior compound eye margin (character 64; figs. 58–60) illustrates the independence of character states. The distinctly sinuate eye of Stichopogoninae (state 2; fig. 60) has evolved from a straight or lightly curved eye margin (state 0; fig. 58) and the presumably intermediate state (sinuate eye in ventral quarter; fig. 59) is found only in some species of the phylogenetically unrelated Asilinae, Dasypogoninae, and Willistonininae. Coding this character as additive would not only increase the length of the most parsimonious cladogram by one step, it would also suggest that the most recent common ancestor of Stichopogoninae first evolved a slightly sinuate eye margin (state 1) and then later, in a second evolutionary event, developed a distinctly sinuate eye margin (state 2). As this phylogenetic analysis is based only on extant taxa and actual phenotypical observations, there is no justification to propose a particular sequence of evolutionary transformations a priori. To impose a particular additivity (or order) on character states would assume that every character state evolves with the same evolutionary rate (homogeneity assumption of molecular data) and evolutionary direction no matter where they are optimized on the cladogram. In some instances, e.g., fusion of labella to prementum (character 23; figs. 31, 36), the transformation series from state 0 to state 1 and finally to state 2 is optimized as the most parsimonious solution, but in other instances (see table 8) such a transformation has not been observed in light of the congruence test with all other included characters.

Phylogenetic Placement of the Two Oldest Known Fossil Robber Flies

The two oldest extinct robber flies, i.e., Araripogon axelrodi Grimaldi, 1990, from the Crato Formation of Brazil and an undescribed species from New Jersey amber (Grimaldi and Cumming, 1999), date to the late Cretaceous and provide a minimum age of Asilidae (see Introduction). Both species have not yet been placed in any higher-level taxon within Asilidae although Grimaldi and Cumming (1999) suggested that the New Jersey amber species shows similarities to extant species of Leptogastrinae. Based on the morphological study of the present phylogenetic hypothesis a placement of these two species in higher-level taxa is possible. Apioceridae and Mydidae, as the sister taxon to Asilidae, and Laphriinae and Asilinae + Ommatiinae, as the two basalmost clades of Asilidae, exhibit closed wing cells r₁ and cup (among other closed wing cells). In contrast, all wing cells are open in Araripogon axelrodi and therefore this species can be placed at the root of clade C (see fig. 121). A more precise placement within this clade is currently not possible as additional information from other morphological characters is needed. D. Grimaldi (personal commun.) has located additional fossil fragments of this species and more character systems can be examined in the future to help place Araripogon axelrodi.

The undescribed species from New Jersey amber can be placed at the root of Leptogastrinae as has been suggested by Grimaldi and Cumming (1999) (see fig. 121). Leptogastrinae, as here delimited, includes the genus Acronyches, which shares characters with the Leptogastrinae sensu stricto, but still retains a few plesiomorphic features. Of the apomorphic character states delimiting Leptogastrinae, the following are present in the extinct species: seta-like spicules on hypopharynx spaced far apart (33; fig. 43), ocellar setae absent (62), presutural dorsocentral setae absent (92). Additional characters supporting the placement are the reduced alula, the few mystacal setae, and the shape of the postpedicel. An autapomorphy of Leptogastrinae sensu previous authors, i.e., anteriorly and medially extended postpronotal lobes (72; fig. 65), is neither present in the fossil nor in Acronyches, which exhibit the plesiomorphic state, corroborating the placement of this extinct species at the root of Leptogastrinae. Another character supporting this particular position is the shape of the frons. The extinct species exhibits a parallel-sided frons as also found in Leptogastrinae sensu previous authors (see, e.g.; fig. 28) whereas in Acronyches the frons is markedly approximating at its dorsal margin. This can be taken as evidence that the extinct species is not more closely related to Acronyches (based on the shape of the postpronotal lobes) than to other Leptogastrinae sensu previous authors.

Based on these two fossil species the minimum age for a taxon composed of clade C and its sister group Asilinae + Ommatiinae can be postulated to be at least 110 Million years (fig. 121). The minimum age for a clade composed of Leptogastrinae and its sister taxon Trigonomiminae can be postulated to be at least 90 Million years (fig. 121).

Sister Group to the Asilidae

As has been summarized in the introduction, there is no previous agreement regarding which taxon represents the sister group to Asilidae within Asiloidea. The present analysis, which includes 16 Asiloidea outgroup taxa, supports a sister-group relationship between Asilidae + (Apioceridae + Mydidae) (figs. 118Article Foldout–Figure 119Article FoldoutFigure 120121). Representatives of these three taxa share the autapomorphic presence of a lateral depression on the prothoracic coxae (character 104). This depression is absent only in Leptogastrinae among Asilidae, and Pseudonomoneura hirta among Mydidae. Additional apomorphic character states supporting the monophyly of this taxon are: clypeal-facial margin forming a distinct ridge (10; figs. 27–31), sensory pit in distal maxillary palpomere absent (12), propresternum square (76; fig. 66), upper calypter small (129), female spermathecal reservoir not differentiated from spermathecal ducts (same diameter) (180; fig. 100). The monophyly of a taxon composed of these three taxa is also supported by the presence of muscles M5, M33, and M38 in the male terminalia (Ovtshinnikova, 1989; Ovtshinnikova and Yeates, 1998) although these characters are uninformative within this clade. I believe that this sister-group relationship is well supported and it has been independently postulated by Yeates and Irwin (1996) in a morphological phylogenetic analysis focused on Apioceridae.

Hennig (1973), Wood (1981), Woodley (1989), and Yeates and Irwin (1996) discussed the depressed vertex in regard to these three taxa and state that it could possibly be informative for deciphering the sister-group relationships. In the present analysis, I divide the depressed vertex into two character states (see character 3) and a sharply depressed vertex of an angle of about 90° is optimized at the root of Asilidae and found in all Asilinae, Laphriinae, and Ommatiinae as well as additional robber flies in clade C (figs. 28, 120). A slightly depressed vertex is found in about half of the species of taxon C as well as among Mydidae with the exception of Neorhaphiomidas and Rhaphiomidas (figs. 27, 30). In these two Mydidae genera, Apioceridae, and all remaining outgroup taxa, the vertex is at the same level as the median compound eye margin. Therefore, the most parsimonious explanation is that the slightly depressed vertex found in some Asilidae of clade C and some Mydidae of a less inclusive clade originated independently as has been suggested previously by Woodley (1989).

Apioceridae and Mydidae form a clade supported by two autapomorphies, i.e., postgenae with medial projection (16) and anterior ocellus separate and situated anteriorly (60; fig. 30), and additional apomorphic character states (54: 5, 55: 1, 116: 1, 124: 0, 127: 2, 145: 0, 147: 0, 174: 1). The single dorsal longitudinal ridge on the pulvilli (see character 124) could be interpreted as another autapomorphy of Apioceridae + Mydidae as suggested by Yeates and Irwin (1996), as it is only found outside of this taxon in Megapodina (Dasypogoninae) and this is certainly an independent reduction from two ridges to one. Woodley (1989) and Yeates and Irwin (1996) found that wing veins R₅ and M₁ (see characters 145 and 147) terminate anterior to the wing apex in Apioceridae and Mydidae (in some Mydidae they do not reach the wing margin at all) and this development is postulated to be an autapomorphy. This is corroborated in the present analysis as this arrangement of veins is only found in the phylogenetically unrelated Nemestrinidae and a few Asilinae species. Woodley (1989), Yeates and Irwin (1996), and Yeates (2002) postulated the supernumerary rectal papillae as an autapomorphy of this clade. Although it appears as if it could be an important character, it has not been examined in many species. Future studies, including more Apioceridae, Mydidae, but also Asilidae, are necessary. Other autapomorphies postulated by Yeates and Irwin (1996) for the clade Apioceridae + Mydidae are: postpedicel ovate, empodium absent, R₄ not reaching costa, male epandrium divided into two halves, female T9 with dorsal ridge, female lateral margin of T9 joining ventrally (not investigated here). Characters that are included in the present analysis and corroborated as apomorphic character states for Apioceridae + Mydidae are: (1) empodium minute or entirely absent (127, also found among several Asilidae clades), (2) female T9 with dorsal ridge (171), (3) divided epandrium (197, also found among many Asilidae with the notable exception of all Laphriinae and some additional taxa). R₄ is found here to reach the wing margin in the genera Apiocera and Rhaphiomidas, but not in the remaining Mydidae, therefore suggesting that this character is autapomorphic for a less inclusive clade (see character 143). The phylogenetic relationships between Apiocera, Neorhaphiomidas, Rhaphiomidas, and remaining Mydidae and family taxon delimitation proposed by Yeates and Irwin (1996) are corroborated here. Apioceridae is supported as monophyletic by the following apomorphic character states: distal maxillary palpomere laterally compressed and leaflike (13), stipites of maxillae entirely fused medially (14), prosternum and proepisternum not fused and prosternum triangular with pointed tip (74), male dorsal aedeagal sheath short and sperm sac entirely free (215). Mydidae is characterized by a single autapomorphy, i.e., male gonostyli absent (205), and a number of apomorphic character states: maxillary palpus one-segmented (11), maxillae shorter than labrum (27), maxillae tapering gradually toward tip (28), apical “seta-like” sensory element apically in cavity on postpedicel (57), postpronotal lobes only partly fused to scutum and lobes elevated above scutum posteriorly (71; fig. 63), cervical sclerite extending medially to cover sensory area on propresternum (77), cell d closed by M₂ and M₃ and these veins not aligned (137), male hypandrium reduced or not differentiated (199), male hypandrium and gonocoxites form gonocoxite-hypandrial complex (202), male lateral ejaculatory process absent (213).